lighthouse/beacon_node/beacon_chain/src/beacon_chain.rs

5048 lines
206 KiB
Rust
Raw Normal View History

Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
use crate::attestation_verification::{
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
batch_verify_aggregated_attestations, batch_verify_unaggregated_attestations,
Error as AttestationError, VerifiedAggregatedAttestation, VerifiedAttestation,
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
VerifiedUnaggregatedAttestation,
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
};
use crate::attester_cache::{AttesterCache, AttesterCacheKey};
use crate::beacon_proposer_cache::compute_proposer_duties_from_head;
Advance state to next slot after importing block (#2174) ## Issue Addressed NA ## Proposed Changes Add an optimization to perform `per_slot_processing` from the *leading-edge* of block processing to the *trailing-edge*. Ultimately, this allows us to import the block at slot `n` faster because we used the tail-end of slot `n - 1` to perform `per_slot_processing`. Additionally, add a "block proposer cache" which allows us to cache the block proposer for some epoch. Since we're now doing trailing-edge `per_slot_processing`, we can prime this cache with the values for the next epoch before those blocks arrive (assuming those blocks don't have some weird forking). There were several ancillary changes required to achieve this: - Remove the `state_root` field of `BeaconSnapshot`, since there's no need to know it on a `pre_state` and in all other cases we can just read it from `block.state_root()`. - This caused some "dust" changes of `snapshot.beacon_state_root` to `snapshot.beacon_state_root()`, where the `BeaconSnapshot::beacon_state_root()` func just reads the state root from the block. - Rename `types::ShuffingId` to `AttestationShufflingId`. I originally did this because I added a `ProposerShufflingId` struct which turned out to be not so useful. I thought this new name was more descriptive so I kept it. - Address https://github.com/ethereum/eth2.0-specs/pull/2196 - Add a debug log when we get a block with an unknown parent. There was previously no logging around this case. - Add a function to `BeaconState` to compute all proposers for an epoch without re-computing the active indices for each slot. ## Additional Info - ~~Blocked on #2173~~ - ~~Blocked on #2179~~ That PR was wrapped into this PR. - There's potentially some places where we could avoid computing the proposer indices in `per_block_processing` but I haven't done this here. These would be an optimization beyond the issue at hand (improving block propagation times) and I think this PR is already doing enough. We can come back for that later. ## TODO - [x] Tidy, improve comments. - [x] ~~Try avoid computing proposer index in `per_block_processing`?~~
2021-02-15 07:17:52 +00:00
use crate::beacon_proposer_cache::BeaconProposerCache;
use crate::block_times_cache::BlockTimesCache;
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
use crate::block_verification::{
check_block_is_finalized_descendant, check_block_relevancy, get_block_root,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
signature_verify_chain_segment, BlockError, ExecutionPendingBlock, GossipVerifiedBlock,
IntoExecutionPendingBlock, PayloadVerificationOutcome, POS_PANDA_BANNER,
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
};
use crate::chain_config::ChainConfig;
Add early attester cache (#2872) ## Issue Addressed NA ## Proposed Changes Introduces a cache to attestation to produce atop blocks which will become the head, but are not fully imported (e.g., not inserted into the database). Whilst attesting to a block before it's imported is rather easy, if we're going to produce that attestation then we also need to be able to: 1. Verify that attestation. 1. Respond to RPC requests for the `beacon_block_root`. Attestation verification (1) is *partially* covered. Since we prime the shuffling cache before we insert the block into the early attester cache, we should be fine for all typical use-cases. However, it is possible that the cache is washed out before we've managed to insert the state into the database and then attestation verification will fail with a "missing beacon state"-type error. Providing the block via RPC (2) is also partially covered, since we'll check the database *and* the early attester cache when responding a blocks-by-root request. However, we'll still omit the block from blocks-by-range requests (until the block lands in the DB). I *think* this is fine, since there's no guarantee that we return all blocks for those responses. Another important consideration is whether or not the *parent* of the early attester block is available in the databse. If it were not, we might fail to respond to blocks-by-root request that are iterating backwards to collect a chain of blocks. I argue that *we will always have the parent of the early attester block in the database.* This is because we are holding the fork-choice write-lock when inserting the block into the early attester cache and we do not drop that until the block is in the database.
2022-01-11 01:35:55 +00:00
use crate::early_attester_cache::EarlyAttesterCache;
Process network attestations (#966) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Remove decompressed member from pubkey bytes * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Change `get_attesting_indices` to use Vec * Fix failing test * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Improve tests * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. * Allow for clock disparity * Ensure errors are returned during batch processing * Add block gossip verification * Connect attestation processing to beacon chain * Optimistically subscribe to subnets on the same slot Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-27 03:10:56 +00:00
use crate::errors::{BeaconChainError as Error, BlockProductionError};
2019-09-03 05:52:25 +00:00
use crate::eth1_chain::{Eth1Chain, Eth1ChainBackend};
Deposit Cache Finalization & Fast WS Sync (#2915) ## Summary The deposit cache now has the ability to finalize deposits. This will cause it to drop unneeded deposit logs and hashes in the deposit Merkle tree that are no longer required to construct deposit proofs. The cache is finalized whenever the latest finalized checkpoint has a new `Eth1Data` with all deposits imported. This has three benefits: 1. Improves the speed of constructing Merkle proofs for deposits as we can just replay deposits since the last finalized checkpoint instead of all historical deposits when re-constructing the Merkle tree. 2. Significantly faster weak subjectivity sync as the deposit cache can be transferred to the newly syncing node in compressed form. The Merkle tree that stores `N` finalized deposits requires a maximum of `log2(N)` hashes. The newly syncing node then only needs to download deposits since the last finalized checkpoint to have a full tree. 3. Future proofing in preparation for [EIP-4444](https://eips.ethereum.org/EIPS/eip-4444) as execution nodes will no longer be required to store logs permanently so we won't always have all historical logs available to us. ## More Details Image to illustrate how the deposit contract merkle tree evolves and finalizes along with the resulting `DepositTreeSnapshot` ![image](https://user-images.githubusercontent.com/37123614/151465302-5fc56284-8a69-4998-b20e-45db3934ac70.png) ## Other Considerations I've changed the structure of the `SszDepositCache` so once you load & save your database from this version of lighthouse, you will no longer be able to load it from older versions. Co-authored-by: ethDreamer <37123614+ethDreamer@users.noreply.github.com>
2022-10-30 04:04:24 +00:00
use crate::eth1_finalization_cache::{Eth1FinalizationCache, Eth1FinalizationData};
use crate::events::ServerSentEventHandler;
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
use crate::execution_payload::{get_execution_payload, PreparePayloadHandle};
Run fork choice before block proposal (#3168) ## Issue Addressed Upcoming spec change https://github.com/ethereum/consensus-specs/pull/2878 ## Proposed Changes 1. Run fork choice at the start of every slot, and wait for this run to complete before proposing a block. 2. As an optimisation, also run fork choice 3/4 of the way through the slot (at 9s), _dequeueing attestations for the next slot_. 3. Remove the fork choice run from the state advance timer that occurred before advancing the state. ## Additional Info ### Block Proposal Accuracy This change makes us more likely to propose on top of the correct head in the presence of re-orgs with proposer boost in play. The main scenario that this change is designed to address is described in the linked spec issue. ### Attestation Accuracy This change _also_ makes us more likely to attest to the correct head. Currently in the case of a skipped slot at `slot` we only run fork choice 9s into `slot - 1`. This means the attestations from `slot - 1` aren't taken into consideration, and any boost applied to the block from `slot - 1` is not removed (it should be). In the language of the linked spec issue, this means we are liable to attest to C, even when the majority voting weight has already caused a re-org to B. ### Why remove the call before the state advance? If we've run fork choice at the start of the slot then it has already dequeued all the attestations from the previous slot, which are the only ones eligible to influence the head in the current slot. Running fork choice again is unnecessary (unless we run it for the next slot and try to pre-empt a re-org, but I don't currently think this is a great idea). ### Performance Based on Prater testing this adds about 5-25ms of runtime to block proposal times, which are 500-1000ms on average (and spike to 5s+ sometimes due to state handling issues :cry: ). I believe this is a small enough penalty to enable it by default, with the option to disable it via the new flag `--fork-choice-before-proposal-timeout 0`. Upcoming work on block packing and state representation will also reduce block production times in general, while removing the spikes. ### Implementation Fork choice gets invoked at the start of the slot via the `per_slot_task` function called from the slot timer. It then uses a condition variable to signal to block production that fork choice has been updated. This is a bit funky, but it seems to work. One downside of the timer-based approach is that it doesn't happen automatically in most of the tests. The test added by this PR has to trigger the run manually.
2022-05-20 05:02:11 +00:00
use crate::fork_choice_signal::{ForkChoiceSignalRx, ForkChoiceSignalTx, ForkChoiceWaitResult};
Prepare for public testnet (#628) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Update to use v0.9.2 version of deposit contract * Add efforts to automate eth1 testnet deployment * Fix lcli testnet deployer * Modify bn CLI to parse eth2_testnet_dir * Progress with account_manager deposit tools * Make account manager submit deposits * Add password option for submitting deposits * Allow custom deposit amount * Add long names to lcli clap * Add password option to lcli deploy command * Fix minor bugs whilst testing * Address Michael's comments * Add refund-deposit-contract to lcli * Use time instead of skip count for denying long skips * Improve logging for eth1 * Fix bug with validator services exiting on error * Drop the block cache after genesis * Modify eth1 testnet config * Improve eth1 logging * Make validator wait until genesis time * Fix bug in eth1 voting * Add more logging to eth1 voting * Handle errors in eth1 http module * Set SECONDS_PER_DAY to sensible minimum * Shorten delay before testnet start * Ensure eth1 block is produced without any votes * Improve eth1 logging * Fix broken tests in eth1 * Tidy code in rest_api * Fix failing test in deposit_contract * Make CLI args more consistent * Change validator/duties endpoint * Add time-based skip slot limiting * Add new error type missed in previous commit * Add log when waiting for genesis * Refactor beacon node CLI * Remove unused dep * Add lcli eth1-genesis command * Fix bug in master merge * Apply clippy lints to beacon node * Add support for YamlConfig in Eth2TestnetDir * Upgrade tesnet deposit contract version * Remove unnecessary logging and correct formatting * Add a hardcoded eth2 testnet config * Ensure http server flag works. Overwrite configs with flags. * Ensure boot nodes are loaded from testnet dir * Fix account manager CLI bugs * Fix bugs with beacon node cli * Allow testnet dir without boot nodes * Write genesis state as SSZ * Remove ---/n from the start of testnet_dir files * Set default libp2p address * Tidy account manager CLI, add logging * Add check to see if testnet dir exists * Apply reviewers suggestions * Add HeadTracker struct * Add fork choice persistence * Shorten slot time for simulator * Add the /beacon/heads API endpoint * Update hardcoded testnet * Add tests for BeaconChain persistence + fix bugs * Extend BeaconChain persistence testing * Ensure chain is finalized b4 persistence tests * Ensure boot_enr.yaml is include in binary * Refactor beacon_chain_sim * Move files about in beacon sim * Update beacon_chain_sim * Fix bug with deposit inclusion * Increase log in genesis service, fix todo * Tidy sim, fix broken rest_api tests * Fix more broken tests * Update testnet * Fix broken rest api test * Tidy account manager CLI * Use tempdir for account manager * Stop hardcoded testnet dir from creating dir * Rename Eth2TestnetDir to Eth2TestnetConfig * Change hardcoded -> hard_coded * Tidy account manager * Add log to account manager * Tidy, ensure head tracker is loaded from disk * Tidy beacon chain builder * Tidy eth1_chain * Adds log support for simulator * Revert "Adds log support for simulator" This reverts commit ec77c66a052350f551db145cf20f213823428dd3. * Adds log support for simulator * Tidy after self-review * Change default log level * Address Michael's delicious PR comments * Fix off-by-one in tests
2019-12-03 04:28:57 +00:00
use crate::head_tracker::HeadTracker;
use crate::historical_blocks::HistoricalBlockError;
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
use crate::migrate::BackgroundMigrator;
use crate::naive_aggregation_pool::{
AggregatedAttestationMap, Error as NaiveAggregationError, NaiveAggregationPool,
SyncContributionAggregateMap,
};
use crate::observed_aggregates::{
Error as AttestationObservationError, ObservedAggregateAttestations, ObservedSyncContributions,
};
use crate::observed_attesters::{
ObservedAggregators, ObservedAttesters, ObservedSyncAggregators, ObservedSyncContributors,
};
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
use crate::observed_block_producers::ObservedBlockProducers;
use crate::observed_operations::{ObservationOutcome, ObservedOperations};
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
use crate::persisted_beacon_chain::{PersistedBeaconChain, DUMMY_CANONICAL_HEAD_BLOCK_ROOT};
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
use crate::persisted_fork_choice::PersistedForkChoice;
use crate::pre_finalization_cache::PreFinalizationBlockCache;
use crate::proposer_prep_service::PAYLOAD_PREPARATION_LOOKAHEAD_FACTOR;
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
use crate::shuffling_cache::{BlockShufflingIds, ShufflingCache};
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
use crate::snapshot_cache::SnapshotCache;
use crate::sync_committee_verification::{
Error as SyncCommitteeError, VerifiedSyncCommitteeMessage, VerifiedSyncContribution,
};
use crate::timeout_rw_lock::TimeoutRwLock;
use crate::validator_monitor::{
get_slot_delay_ms, timestamp_now, ValidatorMonitor,
HISTORIC_EPOCHS as VALIDATOR_MONITOR_HISTORIC_EPOCHS,
};
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
use crate::validator_pubkey_cache::ValidatorPubkeyCache;
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
use crate::BeaconForkChoiceStore;
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
use crate::BeaconSnapshot;
use crate::{metrics, BeaconChainError};
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
use eth2::types::{EventKind, SseBlock, SyncDuty};
use execution_layer::{
BuilderParams, ChainHealth, ExecutionLayer, FailedCondition, PayloadAttributes, PayloadStatus,
};
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
use fork_choice::{
AttestationFromBlock, ExecutionStatus, ForkChoice, ForkchoiceUpdateParameters,
Reset payload statuses when resuming fork choice (#3498) ## Issue Addressed NA ## Proposed Changes This PR is motivated by a recent consensus failure in Geth where it returned `INVALID` for an `VALID` block. Without this PR, the only way to recover is by re-syncing Lighthouse. Whilst ELs "shouldn't have consensus failures", in reality it's something that we can expect from time to time due to the complex nature of Ethereum. Being able to recover easily will help the network recover and EL devs to troubleshoot. The risk introduced with this PR is that genuinely INVALID payloads get a "second chance" at being imported. I believe the DoS risk here is negligible since LH needs to be restarted in order to re-process the payload. Furthermore, there's no reason to think that a well-performing EL will accept a truly invalid payload the second-time-around. ## Additional Info This implementation has the following intricacies: 1. Instead of just resetting *invalid* payloads to optimistic, we'll also reset *valid* payloads. This is an artifact of our existing implementation. 1. We will only reset payload statuses when we detect an invalid payload present in `proto_array` - This helps save us from forgetting that all our blocks are valid in the "best case scenario" where there are no invalid blocks. 1. If we fail to revert the payload statuses we'll log a `CRIT` and just continue with a `proto_array` that *does not* have reverted payload statuses. - The code to revert statuses needs to deal with balances and proposer-boost, so it's a failure point. This is a defensive measure to avoid introducing new show-stopping bugs to LH.
2022-08-29 14:34:41 +00:00
InvalidationOperation, PayloadVerificationStatus, ResetPayloadStatuses,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
};
use futures::channel::mpsc::Sender;
use itertools::process_results;
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
use itertools::Itertools;
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
use operation_pool::{AttestationRef, OperationPool, PersistedOperationPool};
use parking_lot::{Mutex, RwLock};
use proto_array::CountUnrealizedFull;
use safe_arith::SafeArith;
use slasher::Slasher;
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
use slog::{crit, debug, error, info, trace, warn, Logger};
use slot_clock::SlotClock;
use ssz::Encode;
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
use state_processing::{
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
common::{get_attesting_indices_from_state, get_indexed_attestation},
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
per_block_processing,
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
per_block_processing::{
errors::AttestationValidationError, verify_attestation_for_block_inclusion,
VerifySignatures,
},
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
per_slot_processing,
state_advance::{complete_state_advance, partial_state_advance},
BlockSignatureStrategy, ConsensusContext, SigVerifiedOp, VerifyBlockRoot, VerifyOperation,
};
use std::cmp::Ordering;
use std::collections::HashMap;
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
use std::collections::HashSet;
2019-09-09 00:55:15 +00:00
use std::io::prelude::*;
use std::marker::PhantomData;
use std::sync::Arc;
use std::time::{Duration, Instant};
use store::iter::{BlockRootsIterator, ParentRootBlockIterator, StateRootsIterator};
Separate execution payloads in the DB (#3157) ## Proposed Changes Reduce post-merge disk usage by not storing finalized execution payloads in Lighthouse's database. :warning: **This is achieved in a backwards-incompatible way for networks that have already merged** :warning:. Kiln users and shadow fork enjoyers will be unable to downgrade after running the code from this PR. The upgrade migration may take several minutes to run, and can't be aborted after it begins. The main changes are: - New column in the database called `ExecPayload`, keyed by beacon block root. - The `BeaconBlock` column now stores blinded blocks only. - Lots of places that previously used full blocks now use blinded blocks, e.g. analytics APIs, block replay in the DB, etc. - On finalization: - `prune_abanonded_forks` deletes non-canonical payloads whilst deleting non-canonical blocks. - `migrate_db` deletes finalized canonical payloads whilst deleting finalized states. - Conversions between blinded and full blocks are implemented in a compositional way, duplicating some work from Sean's PR #3134. - The execution layer has a new `get_payload_by_block_hash` method that reconstructs a payload using the EE's `eth_getBlockByHash` call. - I've tested manually that it works on Kiln, using Geth and Nethermind. - This isn't necessarily the most efficient method, and new engine APIs are being discussed to improve this: https://github.com/ethereum/execution-apis/pull/146. - We're depending on the `ethers` master branch, due to lots of recent changes. We're also using a workaround for https://github.com/gakonst/ethers-rs/issues/1134. - Payload reconstruction is used in the HTTP API via `BeaconChain::get_block`, which is now `async`. Due to the `async` fn, the `blocking_json` wrapper has been removed. - Payload reconstruction is used in network RPC to serve blocks-by-{root,range} responses. Here the `async` adjustment is messier, although I think I've managed to come up with a reasonable compromise: the handlers take the `SendOnDrop` by value so that they can drop it on _task completion_ (after the `fn` returns). Still, this is introducing disk reads onto core executor threads, which may have a negative performance impact (thoughts appreciated). ## Additional Info - [x] For performance it would be great to remove the cloning of full blocks when converting them to blinded blocks to write to disk. I'm going to experiment with a `put_block` API that takes the block by value, breaks it into a blinded block and a payload, stores the blinded block, and then re-assembles the full block for the caller. - [x] We should measure the latency of blocks-by-root and blocks-by-range responses. - [x] We should add integration tests that stress the payload reconstruction (basic tests done, issue for more extensive tests: https://github.com/sigp/lighthouse/issues/3159) - [x] We should (manually) test the schema v9 migration from several prior versions, particularly as blocks have changed on disk and some migrations rely on being able to load blocks. Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-05-12 00:42:17 +00:00
use store::{
DatabaseBlock, Error as DBError, HotColdDB, KeyValueStore, KeyValueStoreOp, StoreItem, StoreOp,
};
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
use task_executor::{ShutdownReason, TaskExecutor};
Separate execution payloads in the DB (#3157) ## Proposed Changes Reduce post-merge disk usage by not storing finalized execution payloads in Lighthouse's database. :warning: **This is achieved in a backwards-incompatible way for networks that have already merged** :warning:. Kiln users and shadow fork enjoyers will be unable to downgrade after running the code from this PR. The upgrade migration may take several minutes to run, and can't be aborted after it begins. The main changes are: - New column in the database called `ExecPayload`, keyed by beacon block root. - The `BeaconBlock` column now stores blinded blocks only. - Lots of places that previously used full blocks now use blinded blocks, e.g. analytics APIs, block replay in the DB, etc. - On finalization: - `prune_abanonded_forks` deletes non-canonical payloads whilst deleting non-canonical blocks. - `migrate_db` deletes finalized canonical payloads whilst deleting finalized states. - Conversions between blinded and full blocks are implemented in a compositional way, duplicating some work from Sean's PR #3134. - The execution layer has a new `get_payload_by_block_hash` method that reconstructs a payload using the EE's `eth_getBlockByHash` call. - I've tested manually that it works on Kiln, using Geth and Nethermind. - This isn't necessarily the most efficient method, and new engine APIs are being discussed to improve this: https://github.com/ethereum/execution-apis/pull/146. - We're depending on the `ethers` master branch, due to lots of recent changes. We're also using a workaround for https://github.com/gakonst/ethers-rs/issues/1134. - Payload reconstruction is used in the HTTP API via `BeaconChain::get_block`, which is now `async`. Due to the `async` fn, the `blocking_json` wrapper has been removed. - Payload reconstruction is used in network RPC to serve blocks-by-{root,range} responses. Here the `async` adjustment is messier, although I think I've managed to come up with a reasonable compromise: the handlers take the `SendOnDrop` by value so that they can drop it on _task completion_ (after the `fn` returns). Still, this is introducing disk reads onto core executor threads, which may have a negative performance impact (thoughts appreciated). ## Additional Info - [x] For performance it would be great to remove the cloning of full blocks when converting them to blinded blocks to write to disk. I'm going to experiment with a `put_block` API that takes the block by value, breaks it into a blinded block and a payload, stores the blinded block, and then re-assembles the full block for the caller. - [x] We should measure the latency of blocks-by-root and blocks-by-range responses. - [x] We should add integration tests that stress the payload reconstruction (basic tests done, issue for more extensive tests: https://github.com/sigp/lighthouse/issues/3159) - [x] We should (manually) test the schema v9 migration from several prior versions, particularly as blocks have changed on disk and some migrations rely on being able to load blocks. Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-05-12 00:42:17 +00:00
use tree_hash::TreeHash;
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
use types::beacon_state::CloneConfig;
use types::*;
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub use crate::canonical_head::{CanonicalHead, CanonicalHeadRwLock};
pub use fork_choice::CountUnrealized;
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
pub type ForkChoiceError = fork_choice::Error<crate::ForkChoiceStoreError>;
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
/// Alias to appease clippy.
type HashBlockTuple<E> = (Hash256, Arc<SignedBeaconBlock<E>>);
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
/// The time-out before failure during an operation to take a read/write RwLock on the block
/// processing cache.
pub const BLOCK_PROCESSING_CACHE_LOCK_TIMEOUT: Duration = Duration::from_secs(1);
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
/// The time-out before failure during an operation to take a read/write RwLock on the
/// attestation cache.
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
pub const ATTESTATION_CACHE_LOCK_TIMEOUT: Duration = Duration::from_secs(1);
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
/// The time-out before failure during an operation to take a read/write RwLock on the
/// validator pubkey cache.
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
pub const VALIDATOR_PUBKEY_CACHE_LOCK_TIMEOUT: Duration = Duration::from_secs(1);
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
Deposit Cache Finalization & Fast WS Sync (#2915) ## Summary The deposit cache now has the ability to finalize deposits. This will cause it to drop unneeded deposit logs and hashes in the deposit Merkle tree that are no longer required to construct deposit proofs. The cache is finalized whenever the latest finalized checkpoint has a new `Eth1Data` with all deposits imported. This has three benefits: 1. Improves the speed of constructing Merkle proofs for deposits as we can just replay deposits since the last finalized checkpoint instead of all historical deposits when re-constructing the Merkle tree. 2. Significantly faster weak subjectivity sync as the deposit cache can be transferred to the newly syncing node in compressed form. The Merkle tree that stores `N` finalized deposits requires a maximum of `log2(N)` hashes. The newly syncing node then only needs to download deposits since the last finalized checkpoint to have a full tree. 3. Future proofing in preparation for [EIP-4444](https://eips.ethereum.org/EIPS/eip-4444) as execution nodes will no longer be required to store logs permanently so we won't always have all historical logs available to us. ## More Details Image to illustrate how the deposit contract merkle tree evolves and finalizes along with the resulting `DepositTreeSnapshot` ![image](https://user-images.githubusercontent.com/37123614/151465302-5fc56284-8a69-4998-b20e-45db3934ac70.png) ## Other Considerations I've changed the structure of the `SszDepositCache` so once you load & save your database from this version of lighthouse, you will no longer be able to load it from older versions. Co-authored-by: ethDreamer <37123614+ethDreamer@users.noreply.github.com>
2022-10-30 04:04:24 +00:00
/// The timeout for the eth1 finalization cache
pub const ETH1_FINALIZATION_CACHE_LOCK_TIMEOUT: Duration = Duration::from_millis(200);
// These keys are all zero because they get stored in different columns, see `DBColumn` type.
pub const BEACON_CHAIN_DB_KEY: Hash256 = Hash256::zero();
pub const OP_POOL_DB_KEY: Hash256 = Hash256::zero();
pub const ETH1_CACHE_DB_KEY: Hash256 = Hash256::zero();
pub const FORK_CHOICE_DB_KEY: Hash256 = Hash256::zero();
Add early attester cache (#2872) ## Issue Addressed NA ## Proposed Changes Introduces a cache to attestation to produce atop blocks which will become the head, but are not fully imported (e.g., not inserted into the database). Whilst attesting to a block before it's imported is rather easy, if we're going to produce that attestation then we also need to be able to: 1. Verify that attestation. 1. Respond to RPC requests for the `beacon_block_root`. Attestation verification (1) is *partially* covered. Since we prime the shuffling cache before we insert the block into the early attester cache, we should be fine for all typical use-cases. However, it is possible that the cache is washed out before we've managed to insert the state into the database and then attestation verification will fail with a "missing beacon state"-type error. Providing the block via RPC (2) is also partially covered, since we'll check the database *and* the early attester cache when responding a blocks-by-root request. However, we'll still omit the block from blocks-by-range requests (until the block lands in the DB). I *think* this is fine, since there's no guarantee that we return all blocks for those responses. Another important consideration is whether or not the *parent* of the early attester block is available in the databse. If it were not, we might fail to respond to blocks-by-root request that are iterating backwards to collect a chain of blocks. I argue that *we will always have the parent of the early attester block in the database.* This is because we are holding the fork-choice write-lock when inserting the block into the early attester cache and we do not drop that until the block is in the database.
2022-01-11 01:35:55 +00:00
/// Defines how old a block can be before it's no longer a candidate for the early attester cache.
const EARLY_ATTESTER_CACHE_HISTORIC_SLOTS: u64 = 4;
/// Defines a distance between the head block slot and the current slot.
///
/// If the head block is older than this value, don't bother preparing beacon proposers.
const PREPARE_PROPOSER_HISTORIC_EPOCHS: u64 = 4;
Avoid parallel fork choice runs during sync (#3217) ## Issue Addressed Fixes an issue that @paulhauner found with the v2.3.0 release candidate whereby the fork choice runs introduced by #3168 tripped over each other during sync: ``` May 24 23:06:40.542 WARN Error signalling fork choice waiter slot: 3884129, error: ForkChoiceSignalOutOfOrder { current: Slot(3884131), latest: Slot(3884129) }, service: beacon ``` This can occur because fork choice is called from the state advance _and_ the per-slot task. When one of these runs takes a long time it can end up finishing after a run from a later slot, tripping the error above. The problem is resolved by not running either of these fork choice calls during sync. Additionally, these parallel fork choice runs were causing issues in the database: ``` May 24 07:49:05.098 WARN Found a chain that should already have been pruned, head_slot: 92925, head_block_root: 0xa76c7bf1b98e54ed4b0d8686efcfdf853484e6c2a4c67e91cbf19e5ad1f96b17, service: beacon May 24 07:49:05.101 WARN Database migration failed error: HotColdDBError(FreezeSlotError { current_split_slot: Slot(92608), proposed_split_slot: Slot(92576) }), service: beacon ``` In this case, two fork choice calls triggering the finalization processing were being processed out of order due to differences in their processing time, causing the background migrator to try to advance finalization _backwards_ :flushed:. Removing the parallel fork choice runs from sync effectively addresses the issue, because these runs are most likely to have different finalized checkpoints (because of the speed at which fork choice advances during sync). In theory it's still possible to process updates out of order if any other fork choice runs end up completing out of order, but this should be much less common. Fixing out of order fork choice runs in general is difficult as it requires architectural changes like serialising fork choice updates through a single thread, or locking fork choice along with the head when it is mutated (https://github.com/sigp/lighthouse/pull/3175). ## Proposed Changes * Don't run per-slot fork choice during sync (if head is older than 4 slots) * Don't run state-advance fork choice during sync (if head is older than 4 slots) * Check for monotonic finalization updates in the background migrator. This is a good defensive check to have, and I'm not sure why we didn't have it before (we may have had it and wrongly removed it).
2022-05-25 03:27:30 +00:00
/// If the head is more than `MAX_PER_SLOT_FORK_CHOICE_DISTANCE` slots behind the wall-clock slot, DO NOT
/// run the per-slot tasks (primarily fork choice).
///
/// This prevents unnecessary work during sync.
///
/// The value is set to 256 since this would be just over one slot (12.8s) when syncing at
/// 20 slots/second. Having a single fork-choice run interrupt syncing would have very little
/// impact whilst having 8 epochs without a block is a comfortable grace period.
const MAX_PER_SLOT_FORK_CHOICE_DISTANCE: u64 = 256;
Avoid parallel fork choice runs during sync (#3217) ## Issue Addressed Fixes an issue that @paulhauner found with the v2.3.0 release candidate whereby the fork choice runs introduced by #3168 tripped over each other during sync: ``` May 24 23:06:40.542 WARN Error signalling fork choice waiter slot: 3884129, error: ForkChoiceSignalOutOfOrder { current: Slot(3884131), latest: Slot(3884129) }, service: beacon ``` This can occur because fork choice is called from the state advance _and_ the per-slot task. When one of these runs takes a long time it can end up finishing after a run from a later slot, tripping the error above. The problem is resolved by not running either of these fork choice calls during sync. Additionally, these parallel fork choice runs were causing issues in the database: ``` May 24 07:49:05.098 WARN Found a chain that should already have been pruned, head_slot: 92925, head_block_root: 0xa76c7bf1b98e54ed4b0d8686efcfdf853484e6c2a4c67e91cbf19e5ad1f96b17, service: beacon May 24 07:49:05.101 WARN Database migration failed error: HotColdDBError(FreezeSlotError { current_split_slot: Slot(92608), proposed_split_slot: Slot(92576) }), service: beacon ``` In this case, two fork choice calls triggering the finalization processing were being processed out of order due to differences in their processing time, causing the background migrator to try to advance finalization _backwards_ :flushed:. Removing the parallel fork choice runs from sync effectively addresses the issue, because these runs are most likely to have different finalized checkpoints (because of the speed at which fork choice advances during sync). In theory it's still possible to process updates out of order if any other fork choice runs end up completing out of order, but this should be much less common. Fixing out of order fork choice runs in general is difficult as it requires architectural changes like serialising fork choice updates through a single thread, or locking fork choice along with the head when it is mutated (https://github.com/sigp/lighthouse/pull/3175). ## Proposed Changes * Don't run per-slot fork choice during sync (if head is older than 4 slots) * Don't run state-advance fork choice during sync (if head is older than 4 slots) * Check for monotonic finalization updates in the background migrator. This is a good defensive check to have, and I'm not sure why we didn't have it before (we may have had it and wrongly removed it).
2022-05-25 03:27:30 +00:00
/// Reported to the user when the justified block has an invalid execution payload.
pub const INVALID_JUSTIFIED_PAYLOAD_SHUTDOWN_REASON: &str =
"Justified block has an invalid execution payload.";
pub const INVALID_FINALIZED_MERGE_TRANSITION_BLOCK_SHUTDOWN_REASON: &str =
"Finalized merge transition block is invalid.";
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
/// Defines the behaviour when a block/block-root for a skipped slot is requested.
pub enum WhenSlotSkipped {
/// If the slot is a skip slot, return `None`.
///
/// This is how the HTTP API behaves.
None,
/// If the slot it a skip slot, return the previous non-skipped block.
///
/// This is generally how the specification behaves.
Prev,
}
/// The result of a chain segment processing.
pub enum ChainSegmentResult<T: EthSpec> {
/// Processing this chain segment finished successfully.
Successful { imported_blocks: usize },
/// There was an error processing this chain segment. Before the error, some blocks could
/// have been imported.
Failed {
imported_blocks: usize,
error: BlockError<T>,
},
}
/// Configure the signature verification of produced blocks.
pub enum ProduceBlockVerification {
VerifyRandao,
NoVerification,
}
2020-06-19 01:18:27 +00:00
/// The accepted clock drift for nodes gossiping blocks and attestations. See:
Re-merge updated block processing to v0.2.0 (#962) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Remove decompressed member from pubkey bytes * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Change `get_attesting_indices` to use Vec * Fix failing test * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Improve tests * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. * Add sync-speed metric (#898) * Add PH & MS slot clock changes * Account for genesis time * Use checked mul * Account for genesis slot * Change API * Allow for clock disparity * Refactor "duration to..." functions * Ensure errors are returned during batch processing Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-26 06:34:06 +00:00
///
2020-06-19 01:18:27 +00:00
/// https://github.com/ethereum/eth2.0-specs/blob/v0.12.1/specs/phase0/p2p-interface.md#configuration
Re-merge updated block processing to v0.2.0 (#962) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Remove decompressed member from pubkey bytes * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Change `get_attesting_indices` to use Vec * Fix failing test * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Improve tests * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. * Add sync-speed metric (#898) * Add PH & MS slot clock changes * Account for genesis time * Use checked mul * Account for genesis slot * Change API * Allow for clock disparity * Refactor "duration to..." functions * Ensure errors are returned during batch processing Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-26 06:34:06 +00:00
pub const MAXIMUM_GOSSIP_CLOCK_DISPARITY: Duration = Duration::from_millis(500);
Attestation processing (#497) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Fix minor PR comments * Remove duplicated attestation finalization check * Remove awkward `let` statement
2019-08-14 00:55:24 +00:00
#[derive(Debug, PartialEq)]
pub enum AttestationProcessingOutcome {
Processed,
Validator client refactor (#618) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Address Michael's comments
2019-11-25 04:48:24 +00:00
EmptyAggregationBitfield,
Attestation processing (#497) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Fix minor PR comments * Remove duplicated attestation finalization check * Remove awkward `let` statement
2019-08-14 00:55:24 +00:00
UnknownHeadBlock {
beacon_block_root: Hash256,
},
/// The attestation is attesting to a state that is later than itself. (Viz., attesting to the
/// future).
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
AttestsToFutureBlock {
block: Slot,
Attestation processing (#497) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Fix minor PR comments * Remove duplicated attestation finalization check * Remove awkward `let` statement
2019-08-14 00:55:24 +00:00
attestation: Slot,
},
/// The slot is finalized, no need to import.
FinalizedSlot {
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
attestation: Slot,
finalized: Slot,
},
FutureEpoch {
attestation_epoch: Epoch,
current_epoch: Epoch,
},
PastEpoch {
attestation_epoch: Epoch,
current_epoch: Epoch,
},
BadTargetEpoch,
UnknownTargetRoot(Hash256),
InvalidSignature,
NoCommitteeForSlotAndIndex {
slot: Slot,
index: CommitteeIndex,
Attestation processing (#497) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Fix minor PR comments * Remove duplicated attestation finalization check * Remove awkward `let` statement
2019-08-14 00:55:24 +00:00
},
Invalid(AttestationValidationError),
}
Faster attestation production (#838) * Start adding interop genesis state to lcli * Use more efficient method to generate genesis state * Remove duplicate int_to_bytes32 * Add lcli command to change state genesis time * Add option to allow VC to start with unsynced BN * Set VC to do parallel key loading * Don't default to dummy eth1 backend * Add endpoint to dump operation pool * Add metrics for op pool * Remove state clone for slot notifier * Add mem size approximation for tree hash cache * Avoid cloning tree hash when getting head * Avoid cloning tree hash when getting head * Add working arena-based cached tree hash * Add another benchmark * Add pre-allocation for caches * Make cache nullable * Fix bugs in cache tree hash * Add validator tree hash optimization * Optimize hash_concat * Make hash32_concat return fixed-len array * Fix failing API tests * Add new beacon state cache struct * Add validator-specific cache * Separate list and values arenas * Add parallel validator registry hashing * Remove MultiTreeHashCache * Remove cached tree hash macro * Fix failing tree hash test * Address Michael's comments * Add CachedTreeHash impl for ef tests * Fix messy merge conflict * Optimize attestation production * Add first basic optimizations * Fix SlotOutOfBounds error * Resolved missed merge conflicts * Fix another missed merge conflict * Fix more merge conflict issues * Add `StateSkipConfig` * Fix test compile errors * Add failing test * Fix bug, make tests pass * Add comment * Delete unused function * Replace deleted comment
2020-03-04 06:10:22 +00:00
/// Defines how a `BeaconState` should be "skipped" through skip-slots.
pub enum StateSkipConfig {
/// Calculate the state root during each skip slot, producing a fully-valid `BeaconState`.
WithStateRoots,
/// Don't calculate the state root at each slot, instead just use the zero hash. This is orders
/// of magnitude faster, however it produces a partially invalid state.
///
/// This state is useful for operations that don't use the state roots; e.g., for calculating
/// the shuffling.
WithoutStateRoots,
}
Bootstrap (#501) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Begin metrics refactor * Move beacon_chain to new metrics structure. * Make metrics not panic if already defined * Use global prometheus gather at rest api * Unify common metric fns into a crate * Add heavy metering to block processing * Remove hypen from prometheus metric name * Add more beacon chain metrics * Add beacon chain persistence metric * Prune op pool on finalization * Add extra prom beacon chain metrics * Prefix BeaconChain metrics with "beacon_" * Add more store metrics * Add basic metrics to libp2p * Add metrics to HTTP server * Remove old `http_server` crate * Update metrics names to be more like standard * Fix broken beacon chain metrics, add slot clock metrics * Add lighthouse_metrics gather fn * Remove http args * Fix wrong state given to op pool prune * Make prom metric names more consistent * Add more metrics, tidy existing metrics * Fix store block read metrics * Tidy attestation metrics * Fix minor PR comments * Fix minor PR comments * Remove duplicated attestation finalization check * Remove awkward `let` statement * Add first attempts at HTTP bootstrap * Add beacon_block methods to rest api * Fix serde for block.body.grafitti * Allow travis failures on beta (see desc) There's a non-backward compatible change in `cargo fmt`. Stable and beta do not agree. * Add network routes to API * Fix rustc warnings * Add best_slot method * Add --bootstrap arg to beacon node * Get bootstrapper working for ENR address * Store intermediate states during block processing * Allow bootstrapper to scrape libp2p address * Update bootstrapper libp2p address finding * Add comments * Tidy API to be more consistent with recent decisions * Address some review comments * Make BeaconChainTypes Send + Sync + 'static * Add `/network/listen_port` API endpoint * Abandon starting the node if libp2p doesn't start * Update bootstrapper for API changes * Remove unnecessary trait bounds
2019-08-23 05:53:53 +00:00
pub trait BeaconChainTypes: Send + Sync + 'static {
type HotStore: store::ItemStore<Self::EthSpec>;
type ColdStore: store::ItemStore<Self::EthSpec>;
type SlotClock: slot_clock::SlotClock;
type Eth1Chain: Eth1ChainBackend<Self::EthSpec>;
type EthSpec: types::EthSpec;
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
/// Used internally to split block production into discrete functions.
struct PartialBeaconBlock<E: EthSpec, Payload> {
state: BeaconState<E>,
slot: Slot,
proposer_index: u64,
parent_root: Hash256,
randao_reveal: Signature,
eth1_data: Eth1Data,
graffiti: Graffiti,
proposer_slashings: Vec<ProposerSlashing>,
attester_slashings: Vec<AttesterSlashing<E>>,
attestations: Vec<Attestation<E>>,
deposits: Vec<Deposit>,
voluntary_exits: Vec<SignedVoluntaryExit>,
sync_aggregate: Option<SyncAggregate<E>>,
prepare_payload_handle: Option<PreparePayloadHandle<Payload>>,
}
pub type BeaconForkChoice<T> = ForkChoice<
BeaconForkChoiceStore<
<T as BeaconChainTypes>::EthSpec,
<T as BeaconChainTypes>::HotStore,
<T as BeaconChainTypes>::ColdStore,
>,
<T as BeaconChainTypes>::EthSpec,
>;
pub type BeaconStore<T> = Arc<
HotColdDB<
<T as BeaconChainTypes>::EthSpec,
<T as BeaconChainTypes>::HotStore,
<T as BeaconChainTypes>::ColdStore,
>,
>;
/// Represents the "Beacon Chain" component of Ethereum 2.0. Allows import of blocks and block
/// operations and chooses a canonical head.
pub struct BeaconChain<T: BeaconChainTypes> {
pub spec: ChainSpec,
/// Configuration for `BeaconChain` runtime behaviour.
pub config: ChainConfig,
/// Persistent storage for blocks, states, etc. Typically an on-disk store, such as LevelDB.
pub store: BeaconStore<T>,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
/// Used for spawning async and blocking tasks.
pub task_executor: TaskExecutor,
/// Database migrator for running background maintenance on the store.
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
pub store_migrator: BackgroundMigrator<T::EthSpec, T::HotStore, T::ColdStore>,
/// Reports the current slot, typically based upon the system clock.
pub slot_clock: T::SlotClock,
/// Stores all operations (e.g., `Attestation`, `Deposit`, etc) that are candidates for
/// inclusion in a block.
pub op_pool: OperationPool<T::EthSpec>,
/// A pool of attestations dedicated to the "naive aggregation strategy" defined in the eth2
/// specs.
///
/// This pool accepts `Attestation` objects that only have one aggregation bit set and provides
/// a method to get an aggregated `Attestation` for some `AttestationData`.
pub naive_aggregation_pool: RwLock<NaiveAggregationPool<AggregatedAttestationMap<T::EthSpec>>>,
/// A pool of `SyncCommitteeContribution` dedicated to the "naive aggregation strategy" defined in the eth2
/// specs.
///
/// This pool accepts `SyncCommitteeContribution` objects that only have one aggregation bit set and provides
/// a method to get an aggregated `SyncCommitteeContribution` for some `SyncCommitteeContributionData`.
pub naive_sync_aggregation_pool:
RwLock<NaiveAggregationPool<SyncContributionAggregateMap<T::EthSpec>>>,
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
/// Contains a store of attestations which have been observed by the beacon chain.
pub(crate) observed_attestations: RwLock<ObservedAggregateAttestations<T::EthSpec>>,
/// Contains a store of sync contributions which have been observed by the beacon chain.
pub(crate) observed_sync_contributions: RwLock<ObservedSyncContributions<T::EthSpec>>,
Ensure doppelganger detects attestations in blocks (#2495) ## Issue Addressed NA ## Proposed Changes When testing our (not-yet-released) Doppelganger implementation, I noticed that we aren't detecting attestations included in blocks (only those on the gossip network). This is because during [block processing](https://github.com/sigp/lighthouse/blob/e8c0d1f19b2736efb83c67a247e0022da5eaa7bb/beacon_node/beacon_chain/src/beacon_chain.rs#L2168) we only update the `observed_attestations` cache with each attestation, but not the `observed_attesters` cache. This is the correct behaviour when we consider the [p2p spec](https://github.com/ethereum/eth2.0-specs/blob/v1.0.1/specs/phase0/p2p-interface.md): > [IGNORE] There has been no other valid attestation seen on an attestation subnet that has an identical attestation.data.target.epoch and participating validator index. We're doing the right thing here and still allowing attestations on gossip that we've seen in a block. However, this doesn't work so nicely for Doppelganger. To resolve this, I've taken the following steps: - Add a `observed_block_attesters` cache. - Rename `observed_attesters` to `observed_gossip_attesters`. ## TODO - [x] Add a test to ensure a validator that's been seen in a block attestation (but not a gossip attestation) returns `true` for `BeaconChain::validator_seen_at_epoch`. - [x] Add a test to ensure `observed_block_attesters` isn't polluted via gossip attestations and vice versa. Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-08-09 02:43:03 +00:00
/// Maintains a record of which validators have been seen to publish gossip attestations in
/// recent epochs.
pub observed_gossip_attesters: RwLock<ObservedAttesters<T::EthSpec>>,
/// Maintains a record of which validators have been seen to have attestations included in
/// blocks in recent epochs.
pub observed_block_attesters: RwLock<ObservedAttesters<T::EthSpec>>,
/// Maintains a record of which validators have been seen sending sync messages in recent epochs.
pub(crate) observed_sync_contributors: RwLock<ObservedSyncContributors<T::EthSpec>>,
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
/// Maintains a record of which validators have been seen to create `SignedAggregateAndProofs`
/// in recent epochs.
Ensure doppelganger detects attestations in blocks (#2495) ## Issue Addressed NA ## Proposed Changes When testing our (not-yet-released) Doppelganger implementation, I noticed that we aren't detecting attestations included in blocks (only those on the gossip network). This is because during [block processing](https://github.com/sigp/lighthouse/blob/e8c0d1f19b2736efb83c67a247e0022da5eaa7bb/beacon_node/beacon_chain/src/beacon_chain.rs#L2168) we only update the `observed_attestations` cache with each attestation, but not the `observed_attesters` cache. This is the correct behaviour when we consider the [p2p spec](https://github.com/ethereum/eth2.0-specs/blob/v1.0.1/specs/phase0/p2p-interface.md): > [IGNORE] There has been no other valid attestation seen on an attestation subnet that has an identical attestation.data.target.epoch and participating validator index. We're doing the right thing here and still allowing attestations on gossip that we've seen in a block. However, this doesn't work so nicely for Doppelganger. To resolve this, I've taken the following steps: - Add a `observed_block_attesters` cache. - Rename `observed_attesters` to `observed_gossip_attesters`. ## TODO - [x] Add a test to ensure a validator that's been seen in a block attestation (but not a gossip attestation) returns `true` for `BeaconChain::validator_seen_at_epoch`. - [x] Add a test to ensure `observed_block_attesters` isn't polluted via gossip attestations and vice versa. Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-08-09 02:43:03 +00:00
pub observed_aggregators: RwLock<ObservedAggregators<T::EthSpec>>,
/// Maintains a record of which validators have been seen to create `SignedContributionAndProofs`
/// in recent epochs.
pub(crate) observed_sync_aggregators: RwLock<ObservedSyncAggregators<T::EthSpec>>,
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
/// Maintains a record of which validators have proposed blocks for each slot.
pub(crate) observed_block_producers: RwLock<ObservedBlockProducers<T::EthSpec>>,
/// Maintains a record of which validators have submitted voluntary exits.
pub(crate) observed_voluntary_exits: Mutex<ObservedOperations<SignedVoluntaryExit, T::EthSpec>>,
/// Maintains a record of which validators we've seen proposer slashings for.
pub(crate) observed_proposer_slashings: Mutex<ObservedOperations<ProposerSlashing, T::EthSpec>>,
/// Maintains a record of which validators we've seen attester slashings for.
pub(crate) observed_attester_slashings:
Mutex<ObservedOperations<AttesterSlashing<T::EthSpec>, T::EthSpec>>,
2019-09-03 05:52:25 +00:00
/// Provides information from the Ethereum 1 (PoW) chain.
pub eth1_chain: Option<Eth1Chain<T::Eth1Chain, T::EthSpec>>,
[Merge] Implement `execution_layer` (#2635) * Checkout serde_utils from rayonism * Make eth1::http functions pub * Add bones of execution_layer * Modify decoding * Expose Transaction, cargo fmt * Add executePayload * Add all minimal spec endpoints * Start adding json rpc wrapper * Finish custom JSON response handler * Switch to new rpc sending method * Add first test * Fix camelCase * Finish adding tests * Begin threading execution layer into BeaconChain * Fix clippy lints * Fix clippy lints * Thread execution layer into ClientBuilder * Add CLI flags * Add block processing methods to ExecutionLayer * Add block_on to execution_layer * Integrate execute_payload * Add extra_data field * Begin implementing payload handle * Send consensus valid/invalid messages * Fix minor type in task_executor * Call forkchoiceUpdated * Add search for TTD block * Thread TTD into execution layer * Allow producing block with execution payload * Add LRU cache for execution blocks * Remove duplicate 0x on ssz_types serialization * Add tests for block getter methods * Add basic block generator impl * Add is_valid_terminal_block to EL * Verify merge block in block_verification * Partially implement --terminal-block-hash-override * Add terminal_block_hash to ChainSpec * Remove Option from terminal_block_hash in EL * Revert merge changes to consensus/fork_choice * Remove commented-out code * Add bones for handling RPC methods on test server * Add first ExecutionLayer tests * Add testing for finding terminal block * Prevent infinite loops * Add insert_merge_block to block gen * Add block gen test for pos blocks * Start adding payloads to block gen * Fix clippy lints * Add execution payload to block gen * Add execute_payload to block_gen * Refactor block gen * Add all routes to mock server * Use Uint256 for base_fee_per_gas * Add working execution chain build * Remove unused var * Revert "Use Uint256 for base_fee_per_gas" This reverts commit 6c88f19ac45db834dd4dbf7a3c6e7242c1c0f735. * Fix base_fee_for_gas Uint256 * Update execute payload handle * Improve testing, fix bugs * Fix default fee-recipient * Fix fee-recipient address (again) * Add check for terminal block, add comments, tidy * Apply suggestions from code review Co-authored-by: realbigsean <seananderson33@GMAIL.com> * Fix is_none on handle Drop * Remove commented-out tests Co-authored-by: realbigsean <seananderson33@GMAIL.com>
2021-09-29 22:14:15 +00:00
/// Interfaces with the execution client.
pub execution_layer: Option<ExecutionLayer<T::EthSpec>>,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
/// Stores information about the canonical head and finalized/justified checkpoints of the
/// chain. Also contains the fork choice struct, for computing the canonical head.
pub canonical_head: CanonicalHead<T>,
/// The root of the genesis block.
pub genesis_block_root: Hash256,
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
/// The root of the genesis state.
pub genesis_state_root: Hash256,
2020-04-08 06:46:37 +00:00
/// The root of the list of genesis validators, used during syncing.
pub genesis_validators_root: Hash256,
Run fork choice before block proposal (#3168) ## Issue Addressed Upcoming spec change https://github.com/ethereum/consensus-specs/pull/2878 ## Proposed Changes 1. Run fork choice at the start of every slot, and wait for this run to complete before proposing a block. 2. As an optimisation, also run fork choice 3/4 of the way through the slot (at 9s), _dequeueing attestations for the next slot_. 3. Remove the fork choice run from the state advance timer that occurred before advancing the state. ## Additional Info ### Block Proposal Accuracy This change makes us more likely to propose on top of the correct head in the presence of re-orgs with proposer boost in play. The main scenario that this change is designed to address is described in the linked spec issue. ### Attestation Accuracy This change _also_ makes us more likely to attest to the correct head. Currently in the case of a skipped slot at `slot` we only run fork choice 9s into `slot - 1`. This means the attestations from `slot - 1` aren't taken into consideration, and any boost applied to the block from `slot - 1` is not removed (it should be). In the language of the linked spec issue, this means we are liable to attest to C, even when the majority voting weight has already caused a re-org to B. ### Why remove the call before the state advance? If we've run fork choice at the start of the slot then it has already dequeued all the attestations from the previous slot, which are the only ones eligible to influence the head in the current slot. Running fork choice again is unnecessary (unless we run it for the next slot and try to pre-empt a re-org, but I don't currently think this is a great idea). ### Performance Based on Prater testing this adds about 5-25ms of runtime to block proposal times, which are 500-1000ms on average (and spike to 5s+ sometimes due to state handling issues :cry: ). I believe this is a small enough penalty to enable it by default, with the option to disable it via the new flag `--fork-choice-before-proposal-timeout 0`. Upcoming work on block packing and state representation will also reduce block production times in general, while removing the spikes. ### Implementation Fork choice gets invoked at the start of the slot via the `per_slot_task` function called from the slot timer. It then uses a condition variable to signal to block production that fork choice has been updated. This is a bit funky, but it seems to work. One downside of the timer-based approach is that it doesn't happen automatically in most of the tests. The test added by this PR has to trigger the run manually.
2022-05-20 05:02:11 +00:00
/// Transmitter used to indicate that slot-start fork choice has completed running.
pub fork_choice_signal_tx: Option<ForkChoiceSignalTx>,
/// Receiver used by block production to wait on slot-start fork choice.
pub fork_choice_signal_rx: Option<ForkChoiceSignalRx>,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
/// The genesis time of this `BeaconChain` (seconds since UNIX epoch).
pub genesis_time: u64,
/// A handler for events generated by the beacon chain. This is only initialized when the
/// HTTP server is enabled.
pub event_handler: Option<ServerSentEventHandler<T::EthSpec>>,
Prepare for public testnet (#628) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Update to use v0.9.2 version of deposit contract * Add efforts to automate eth1 testnet deployment * Fix lcli testnet deployer * Modify bn CLI to parse eth2_testnet_dir * Progress with account_manager deposit tools * Make account manager submit deposits * Add password option for submitting deposits * Allow custom deposit amount * Add long names to lcli clap * Add password option to lcli deploy command * Fix minor bugs whilst testing * Address Michael's comments * Add refund-deposit-contract to lcli * Use time instead of skip count for denying long skips * Improve logging for eth1 * Fix bug with validator services exiting on error * Drop the block cache after genesis * Modify eth1 testnet config * Improve eth1 logging * Make validator wait until genesis time * Fix bug in eth1 voting * Add more logging to eth1 voting * Handle errors in eth1 http module * Set SECONDS_PER_DAY to sensible minimum * Shorten delay before testnet start * Ensure eth1 block is produced without any votes * Improve eth1 logging * Fix broken tests in eth1 * Tidy code in rest_api * Fix failing test in deposit_contract * Make CLI args more consistent * Change validator/duties endpoint * Add time-based skip slot limiting * Add new error type missed in previous commit * Add log when waiting for genesis * Refactor beacon node CLI * Remove unused dep * Add lcli eth1-genesis command * Fix bug in master merge * Apply clippy lints to beacon node * Add support for YamlConfig in Eth2TestnetDir * Upgrade tesnet deposit contract version * Remove unnecessary logging and correct formatting * Add a hardcoded eth2 testnet config * Ensure http server flag works. Overwrite configs with flags. * Ensure boot nodes are loaded from testnet dir * Fix account manager CLI bugs * Fix bugs with beacon node cli * Allow testnet dir without boot nodes * Write genesis state as SSZ * Remove ---/n from the start of testnet_dir files * Set default libp2p address * Tidy account manager CLI, add logging * Add check to see if testnet dir exists * Apply reviewers suggestions * Add HeadTracker struct * Add fork choice persistence * Shorten slot time for simulator * Add the /beacon/heads API endpoint * Update hardcoded testnet * Add tests for BeaconChain persistence + fix bugs * Extend BeaconChain persistence testing * Ensure chain is finalized b4 persistence tests * Ensure boot_enr.yaml is include in binary * Refactor beacon_chain_sim * Move files about in beacon sim * Update beacon_chain_sim * Fix bug with deposit inclusion * Increase log in genesis service, fix todo * Tidy sim, fix broken rest_api tests * Fix more broken tests * Update testnet * Fix broken rest api test * Tidy account manager CLI * Use tempdir for account manager * Stop hardcoded testnet dir from creating dir * Rename Eth2TestnetDir to Eth2TestnetConfig * Change hardcoded -> hard_coded * Tidy account manager * Add log to account manager * Tidy, ensure head tracker is loaded from disk * Tidy beacon chain builder * Tidy eth1_chain * Adds log support for simulator * Revert "Adds log support for simulator" This reverts commit ec77c66a052350f551db145cf20f213823428dd3. * Adds log support for simulator * Tidy after self-review * Change default log level * Address Michael's delicious PR comments * Fix off-by-one in tests
2019-12-03 04:28:57 +00:00
/// Used to track the heads of the beacon chain.
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
pub(crate) head_tracker: Arc<HeadTracker>,
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
/// A cache dedicated to block processing.
pub(crate) snapshot_cache: TimeoutRwLock<SnapshotCache<T::EthSpec>>,
Advance state to next slot after importing block (#2174) ## Issue Addressed NA ## Proposed Changes Add an optimization to perform `per_slot_processing` from the *leading-edge* of block processing to the *trailing-edge*. Ultimately, this allows us to import the block at slot `n` faster because we used the tail-end of slot `n - 1` to perform `per_slot_processing`. Additionally, add a "block proposer cache" which allows us to cache the block proposer for some epoch. Since we're now doing trailing-edge `per_slot_processing`, we can prime this cache with the values for the next epoch before those blocks arrive (assuming those blocks don't have some weird forking). There were several ancillary changes required to achieve this: - Remove the `state_root` field of `BeaconSnapshot`, since there's no need to know it on a `pre_state` and in all other cases we can just read it from `block.state_root()`. - This caused some "dust" changes of `snapshot.beacon_state_root` to `snapshot.beacon_state_root()`, where the `BeaconSnapshot::beacon_state_root()` func just reads the state root from the block. - Rename `types::ShuffingId` to `AttestationShufflingId`. I originally did this because I added a `ProposerShufflingId` struct which turned out to be not so useful. I thought this new name was more descriptive so I kept it. - Address https://github.com/ethereum/eth2.0-specs/pull/2196 - Add a debug log when we get a block with an unknown parent. There was previously no logging around this case. - Add a function to `BeaconState` to compute all proposers for an epoch without re-computing the active indices for each slot. ## Additional Info - ~~Blocked on #2173~~ - ~~Blocked on #2179~~ That PR was wrapped into this PR. - There's potentially some places where we could avoid computing the proposer indices in `per_block_processing` but I haven't done this here. These would be an optimization beyond the issue at hand (improving block propagation times) and I think this PR is already doing enough. We can come back for that later. ## TODO - [x] Tidy, improve comments. - [x] ~~Try avoid computing proposer index in `per_block_processing`?~~
2021-02-15 07:17:52 +00:00
/// Caches the attester shuffling for a given epoch and shuffling key root.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub shuffling_cache: TimeoutRwLock<ShufflingCache>,
Deposit Cache Finalization & Fast WS Sync (#2915) ## Summary The deposit cache now has the ability to finalize deposits. This will cause it to drop unneeded deposit logs and hashes in the deposit Merkle tree that are no longer required to construct deposit proofs. The cache is finalized whenever the latest finalized checkpoint has a new `Eth1Data` with all deposits imported. This has three benefits: 1. Improves the speed of constructing Merkle proofs for deposits as we can just replay deposits since the last finalized checkpoint instead of all historical deposits when re-constructing the Merkle tree. 2. Significantly faster weak subjectivity sync as the deposit cache can be transferred to the newly syncing node in compressed form. The Merkle tree that stores `N` finalized deposits requires a maximum of `log2(N)` hashes. The newly syncing node then only needs to download deposits since the last finalized checkpoint to have a full tree. 3. Future proofing in preparation for [EIP-4444](https://eips.ethereum.org/EIPS/eip-4444) as execution nodes will no longer be required to store logs permanently so we won't always have all historical logs available to us. ## More Details Image to illustrate how the deposit contract merkle tree evolves and finalizes along with the resulting `DepositTreeSnapshot` ![image](https://user-images.githubusercontent.com/37123614/151465302-5fc56284-8a69-4998-b20e-45db3934ac70.png) ## Other Considerations I've changed the structure of the `SszDepositCache` so once you load & save your database from this version of lighthouse, you will no longer be able to load it from older versions. Co-authored-by: ethDreamer <37123614+ethDreamer@users.noreply.github.com>
2022-10-30 04:04:24 +00:00
/// A cache of eth1 deposit data at epoch boundaries for deposit finalization
pub eth1_finalization_cache: TimeoutRwLock<Eth1FinalizationCache>,
Advance state to next slot after importing block (#2174) ## Issue Addressed NA ## Proposed Changes Add an optimization to perform `per_slot_processing` from the *leading-edge* of block processing to the *trailing-edge*. Ultimately, this allows us to import the block at slot `n` faster because we used the tail-end of slot `n - 1` to perform `per_slot_processing`. Additionally, add a "block proposer cache" which allows us to cache the block proposer for some epoch. Since we're now doing trailing-edge `per_slot_processing`, we can prime this cache with the values for the next epoch before those blocks arrive (assuming those blocks don't have some weird forking). There were several ancillary changes required to achieve this: - Remove the `state_root` field of `BeaconSnapshot`, since there's no need to know it on a `pre_state` and in all other cases we can just read it from `block.state_root()`. - This caused some "dust" changes of `snapshot.beacon_state_root` to `snapshot.beacon_state_root()`, where the `BeaconSnapshot::beacon_state_root()` func just reads the state root from the block. - Rename `types::ShuffingId` to `AttestationShufflingId`. I originally did this because I added a `ProposerShufflingId` struct which turned out to be not so useful. I thought this new name was more descriptive so I kept it. - Address https://github.com/ethereum/eth2.0-specs/pull/2196 - Add a debug log when we get a block with an unknown parent. There was previously no logging around this case. - Add a function to `BeaconState` to compute all proposers for an epoch without re-computing the active indices for each slot. ## Additional Info - ~~Blocked on #2173~~ - ~~Blocked on #2179~~ That PR was wrapped into this PR. - There's potentially some places where we could avoid computing the proposer indices in `per_block_processing` but I haven't done this here. These would be an optimization beyond the issue at hand (improving block propagation times) and I think this PR is already doing enough. We can come back for that later. ## TODO - [x] Tidy, improve comments. - [x] ~~Try avoid computing proposer index in `per_block_processing`?~~
2021-02-15 07:17:52 +00:00
/// Caches the beacon block proposer shuffling for a given epoch and shuffling key root.
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
pub beacon_proposer_cache: Mutex<BeaconProposerCache>,
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
/// Caches a map of `validator_index -> validator_pubkey`.
pub(crate) validator_pubkey_cache: TimeoutRwLock<ValidatorPubkeyCache<T>>,
/// A cache used when producing attestations.
pub(crate) attester_cache: Arc<AttesterCache>,
Add early attester cache (#2872) ## Issue Addressed NA ## Proposed Changes Introduces a cache to attestation to produce atop blocks which will become the head, but are not fully imported (e.g., not inserted into the database). Whilst attesting to a block before it's imported is rather easy, if we're going to produce that attestation then we also need to be able to: 1. Verify that attestation. 1. Respond to RPC requests for the `beacon_block_root`. Attestation verification (1) is *partially* covered. Since we prime the shuffling cache before we insert the block into the early attester cache, we should be fine for all typical use-cases. However, it is possible that the cache is washed out before we've managed to insert the state into the database and then attestation verification will fail with a "missing beacon state"-type error. Providing the block via RPC (2) is also partially covered, since we'll check the database *and* the early attester cache when responding a blocks-by-root request. However, we'll still omit the block from blocks-by-range requests (until the block lands in the DB). I *think* this is fine, since there's no guarantee that we return all blocks for those responses. Another important consideration is whether or not the *parent* of the early attester block is available in the databse. If it were not, we might fail to respond to blocks-by-root request that are iterating backwards to collect a chain of blocks. I argue that *we will always have the parent of the early attester block in the database.* This is because we are holding the fork-choice write-lock when inserting the block into the early attester cache and we do not drop that until the block is in the database.
2022-01-11 01:35:55 +00:00
/// A cache used when producing attestations whilst the head block is still being imported.
pub early_attester_cache: EarlyAttesterCache<T::EthSpec>,
/// A cache used to keep track of various block timings.
pub block_times_cache: Arc<RwLock<BlockTimesCache>>,
/// A cache used to track pre-finalization block roots for quick rejection.
pub pre_finalization_block_cache: PreFinalizationBlockCache,
/// Sender given to tasks, so that if they encounter a state in which execution cannot
/// continue they can request that everything shuts down.
pub shutdown_sender: Sender<ShutdownReason>,
Testnet stability (#451) * Change reduced tree for adding weightless node * Add more comments for reduced tree fork choice * Small refactor on reduced tree for readability * Move test_harness forking logic into itself * Add new `AncestorIter` trait to store * Add unfinished tests to fork choice * Make `beacon_state.genesis_block_root` public * Add failing lmd_ghost fork choice tests * Extend fork_choice tests, create failing test * Implement Debug for generic ReducedTree * Add lazy_static to fork choice tests * Add verify_integrity fn to reduced tree * Fix bugs in reduced tree * Ensure all reduced tree tests verify integrity * Slightly alter reduce tree test params * Add (failing) reduced tree test * Fix bug in fork choice Iter ancestors was not working well with skip slots * Put maximum depth for common ancestor search Ensures that we don't search back past the finalized root. * Add basic finalization tests for reduced tree * Change fork choice to use beacon_block_root Previously it was using target_root, which was wrong * Change reduced tree for adding weightless node * Add more comments for reduced tree fork choice * Small refactor on reduced tree for readability * Move test_harness forking logic into itself * Add new `AncestorIter` trait to store * Add unfinished tests to fork choice * Make `beacon_state.genesis_block_root` public * Add failing lmd_ghost fork choice tests * Extend fork_choice tests, create failing test * Implement Debug for generic ReducedTree * Add lazy_static to fork choice tests * Add verify_integrity fn to reduced tree * Fix bugs in reduced tree * Ensure all reduced tree tests verify integrity * Slightly alter reduce tree test params * Add (failing) reduced tree test * Fix bug in fork choice Iter ancestors was not working well with skip slots * Put maximum depth for common ancestor search Ensures that we don't search back past the finalized root. * Add basic finalization tests for reduced tree * Add network dir CLI flag * Simplify "NewSlot" log message * Rename network-dir CLI flag * Change fork choice to use beacon_block_root Previously it was using target_root, which was wrong * Update db dir size for metrics * Change slog to use `FullFormat` logging * Update some comments and log formatting * Add prom gauge for best block root * Only add known target blocks to fork choice * Add finalized and justified root prom metrics * Add CLI flag for setting log level * Add logger to beacon chain * Add debug-level CLI flag to validator * Allow block processing if fork choice fails * Create warn log when there's low libp2p peer count * Minor change to logging * Make ancestor iter return option * Disable fork choice test when !debug_assertions * Fix type, removed code fragment * Tidy some borrow-checker evading * Lower reduced tree random test iterations
2019-07-29 03:45:45 +00:00
/// Logging to CLI, etc.
Eth1 Integration (#542) * Refactor to cache Eth1Data * Fix merge conflicts and minor refactorings * Rename Eth1Cache to Eth1DataCache * Refactor events subscription * Add deposits module to interface with BeaconChain deposits * Remove utils * Rename to types.rs and add trait constraints to Eth1DataFetcher * Confirm to trait constraints. Make Web3DataFetcher cloneable * Make fetcher object member of deposit and eth1_data cache and other fixes * Fix update_cache function * Move fetch_eth1_data to impl block * Fix deposit tests * Create Eth1 object for interfacing with Beacon chain * Add `run` function for running update_cache and subscribe_deposit_logs tasks * Add logging * Run `cargo fmt` and make tests pass * Convert sync functions to async * Add timeouts to web3 functions * Return futures from cache functions * Add failed chaining of futures * Working cache updation * Clean up tests and `update_cache` function * Refactor `get_eth1_data` functions to work with future returning functions * Refactor eth1 `run` function to work with modified `update_cache` api * Minor changes * Add distance parameter to `update_cache` * Fix tests and other minor fixes * Working integration with cache and deposits * Add merkle_tree construction, proof generation and verification code * Add function to construct and fetch Deposits for BeaconNode * Add error handling * Import ssz * Add error handling to eth1 cache and fix minor errors * Run rustfmt * Fix minor bug * Rename Eth1Error and change to Result<T> * Change deposit fetching mechanism from notification based to poll based * Add deposits from eth1 chain in a given range every `x` blocks * Modify `run` function to accommodate changes * Minor fixes * Fix formatting * Initial commit. web3 api working. * Tidied up lib. Add function for fetching logs. * Refactor with `Eth1DataFetcher` trait * Add parsing for deposit contract logs and get_eth1_data function * Add `get_eth1_votes` function * Refactor to cache Eth1Data * Fix merge conflicts and minor refactorings * Rename Eth1Cache to Eth1DataCache * Refactor events subscription * Add deposits module to interface with BeaconChain deposits * Remove utils * Rename to types.rs and add trait constraints to Eth1DataFetcher * Confirm to trait constraints. Make Web3DataFetcher cloneable * Make fetcher object member of deposit and eth1_data cache and other fixes * Fix update_cache function * Move fetch_eth1_data to impl block * Fix deposit tests * Create Eth1 object for interfacing with Beacon chain * Add `run` function for running update_cache and subscribe_deposit_logs tasks * Add logging * Run `cargo fmt` and make tests pass * Convert sync functions to async * Add timeouts to web3 functions * Return futures from cache functions * Add failed chaining of futures * Working cache updation * Clean up tests and `update_cache` function * Refactor `get_eth1_data` functions to work with future returning functions * Refactor eth1 `run` function to work with modified `update_cache` api * Minor changes * Add distance parameter to `update_cache` * Fix tests and other minor fixes * Working integration with cache and deposits * Add merkle_tree construction, proof generation and verification code * Add function to construct and fetch Deposits for BeaconNode * Add error handling * Import ssz * Add error handling to eth1 cache and fix minor errors * Run rustfmt * Fix minor bug * Rename Eth1Error and change to Result<T> * Change deposit fetching mechanism from notification based to poll based * Add deposits from eth1 chain in a given range every `x` blocks * Modify `run` function to accommodate changes * Minor fixes * Fix formatting * Fix merge issue * Refactor with `Config` struct. Remote `ContractConfig` * Rename eth1_chain crate to eth1 * Rename files and read abi file using `fs::read` * Move eth1 to lib * Remove unnecessary mutability constraint * Add `Web3Backend` for returning actual eth1 data * Refactor `get_eth1_votes` to return a Result * Delete `eth1_chain` crate * Return `Result` from `get_deposits` * Fix range of deposits to return to beacon chain * Add `get_block_height_by_hash` trait function * Add naive method for getting `previous_eth1_distance` * Add eth1 config params to main config * Add instructions for setting up eth1 testing environment * Add build script to fetch deposit contract abi * Contract ABI is part of compiled binary * Fix minor bugs * Move docs to lib * Add timeout to config * Remove print statements * Change warn to error * Fix typos * Removed prints in test and get timeout value from config * Fixed error types * Added logging to web3_fetcher * Refactor for modified web3 api * Fix minor stuff * Add build script * Tidy, hide eth1 integration tests behind flag * Add http crate * Add first stages of eth1_test_rig * Fix deposits on test rig * Fix bug with deposit count method * Add block hash getter to http eth1 * Clean eth1 http crate and tests * Add script to start ganache * Adds deposit tree to eth1-http * Extend deposit tree tests * Tidy tests in eth1-http * Add more detail to get block request * Add block cache to eth1-http * Rename deposit tree to deposit cache * Add inital updating to eth1-http * Tidy updater * Fix compile bugs in tests * Adds an Eth1DataCache builder * Reorg eth1-http files * Add (failing) tests for eth1 updater * Rename files, fix bug in eth1-http * Ensure that ganache timestamps are increasing * Fix bugs with getting eth1data ancestors * Improve eth1 testing, fix bugs * Add truncate method to block cache * Add pruning to block cache update process * Add tests for block pruning * Allow for dropping an expired cache. * Add more comments * Add first compiling version of deposit updater * Add common fn for getting range of required blocks * Add passing deposit update test * Improve tests * Fix block pruning bug * Add tests for running two updates at once * Add updater services to eth1 * Add deposit collection to beacon chain * Add incomplete builder experiments * Add first working version of beacon chain builder * Update test harness to new beacon chain type * Rename builder file, tidy * Add first working client builder * Progress further on client builder * Update becaon node binary to use client builder * Ensure release tests compile * Remove old eth1 crate * Add first pass of new lighthouse binary * Fix websocket server startup * Remove old binary code from beacon_node crate * Add first working beacon node tests * Add genesis crate, new eth1 cache_2 * Add Serivce to Eth1Cache * Refactor with general eth1 improvements * Add passing genesis test * Tidy, add comments * Add more comments to eth1 service * Add further eth1 progress * Fix some bugs with genesis * Fix eth1 bugs, make eth1 linking more efficient * Shift logic in genesis service * Add more comments to genesis service * Add gzip, max request values, timeouts to http * Update testnet parameters to suit goerli testnet * Add ability to vary Fork, fix custom spec * Be more explicit about deposit fork version * Start adding beacon chain eth1 option * Add more flexibility to prod client * Further runtime refactoring * Allow for starting from store * Add bootstrapping to client config * Add remote_beacon_node crate * Update eth1 service for more configurability * Update eth1 tests to use less runtimes * Patch issues with tests using too many files * Move dummy eth1 backend flag * Ensure all tests pass * Add ganache-cli to Dockerfile * Use a special docker hub image for testing * Appease clippy * Move validator client into lighthouse binary * Allow starting with dummy eth1 backend * Improve logging * Fix dummy eth1 backend from cli * Add extra testnet command * Ensure consistent spec in beacon node * Update eth1 rig to work on goerli * Tidy lcli, start adding support for yaml config * Add incomplete YamlConfig struct * Remove efforts at YamlConfig * Add incomplete eth1 voting. Blocked on spec issues * Add (untested) first pass at eth1 vote algo * Add tests for winning vote * Add more tests for eth1 chain * Add more eth1 voting tests * Added more eth1 voting testing * Change test name * Add more tests to eth1 chain * Tidy eth1 generics, add more tests * Improve comments * Tidy beacon_node tests * Tidy, rename JsonRpc.. to Caching.. * Tidy voting logic * Tidy builder docs * Add comments, tidy eth1 * Add more comments to eth1 * Fix bug with winning_vote * Add doc comments to the `ClientBuilder` * Remove commented-out code * Improve `ClientBuilder` docs * Add comments to client config * Add decoding test for `ClientConfig` * Remove unused `DepositSet` struct * Tidy `block_cache` * Remove commented out lines * Remove unused code in `eth1` crate * Remove old validator binary `main.rs` * Tidy, fix tests compile error * Add initial tests for get_deposits * Remove dead code in eth1_test_rig * Update TestingDepositBuilder * Add testing for getting eth1 deposits * Fix duplicate rand dep * Remove dead code * Remove accidentally-added files * Fix comment in eth1_genesis_service * Add .gitignore for eth1_test_rig * Fix bug in eth1_genesis_service * Remove dead code from eth2_config * Fix tabs/spaces in root Cargo.toml * Tidy eth1 crate * Allow for re-use of eth1 service after genesis * Update docs for new CLI * Change README gif * Tidy eth1 http module * Tidy eth1 service * Tidy environment crate * Remove unused file * Tidy, add comments * Remove commented-out code * Address majority of Michael's comments * Address other PR comments * Add link to issue alongside TODO
2019-11-15 03:47:51 +00:00
pub(crate) log: Logger,
/// Arbitrary bytes included in the blocks.
pub(crate) graffiti: Graffiti,
/// Optional slasher.
pub slasher: Option<Arc<Slasher<T::EthSpec>>>,
/// Provides monitoring of a set of explicitly defined validators.
pub validator_monitor: RwLock<ValidatorMonitor<T::EthSpec>>,
}
type BeaconBlockAndState<T, Payload> = (BeaconBlock<T, Payload>, BeaconState<T>);
impl<T: BeaconChainTypes> BeaconChain<T> {
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
/// Persists the head tracker and fork choice.
///
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
/// We do it atomically even though no guarantees need to be made about blocks from
/// the head tracker also being present in fork choice.
pub fn persist_head_and_fork_choice(&self) -> Result<(), Error> {
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
let mut batch = vec![];
Prepare for public testnet (#628) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Update to use v0.9.2 version of deposit contract * Add efforts to automate eth1 testnet deployment * Fix lcli testnet deployer * Modify bn CLI to parse eth2_testnet_dir * Progress with account_manager deposit tools * Make account manager submit deposits * Add password option for submitting deposits * Allow custom deposit amount * Add long names to lcli clap * Add password option to lcli deploy command * Fix minor bugs whilst testing * Address Michael's comments * Add refund-deposit-contract to lcli * Use time instead of skip count for denying long skips * Improve logging for eth1 * Fix bug with validator services exiting on error * Drop the block cache after genesis * Modify eth1 testnet config * Improve eth1 logging * Make validator wait until genesis time * Fix bug in eth1 voting * Add more logging to eth1 voting * Handle errors in eth1 http module * Set SECONDS_PER_DAY to sensible minimum * Shorten delay before testnet start * Ensure eth1 block is produced without any votes * Improve eth1 logging * Fix broken tests in eth1 * Tidy code in rest_api * Fix failing test in deposit_contract * Make CLI args more consistent * Change validator/duties endpoint * Add time-based skip slot limiting * Add new error type missed in previous commit * Add log when waiting for genesis * Refactor beacon node CLI * Remove unused dep * Add lcli eth1-genesis command * Fix bug in master merge * Apply clippy lints to beacon node * Add support for YamlConfig in Eth2TestnetDir * Upgrade tesnet deposit contract version * Remove unnecessary logging and correct formatting * Add a hardcoded eth2 testnet config * Ensure http server flag works. Overwrite configs with flags. * Ensure boot nodes are loaded from testnet dir * Fix account manager CLI bugs * Fix bugs with beacon node cli * Allow testnet dir without boot nodes * Write genesis state as SSZ * Remove ---/n from the start of testnet_dir files * Set default libp2p address * Tidy account manager CLI, add logging * Add check to see if testnet dir exists * Apply reviewers suggestions * Add HeadTracker struct * Add fork choice persistence * Shorten slot time for simulator * Add the /beacon/heads API endpoint * Update hardcoded testnet * Add tests for BeaconChain persistence + fix bugs * Extend BeaconChain persistence testing * Ensure chain is finalized b4 persistence tests * Ensure boot_enr.yaml is include in binary * Refactor beacon_chain_sim * Move files about in beacon sim * Update beacon_chain_sim * Fix bug with deposit inclusion * Increase log in genesis service, fix todo * Tidy sim, fix broken rest_api tests * Fix more broken tests * Update testnet * Fix broken rest api test * Tidy account manager CLI * Use tempdir for account manager * Stop hardcoded testnet dir from creating dir * Rename Eth2TestnetDir to Eth2TestnetConfig * Change hardcoded -> hard_coded * Tidy account manager * Add log to account manager * Tidy, ensure head tracker is loaded from disk * Tidy beacon chain builder * Tidy eth1_chain * Adds log support for simulator * Revert "Adds log support for simulator" This reverts commit ec77c66a052350f551db145cf20f213823428dd3. * Adds log support for simulator * Tidy after self-review * Change default log level * Address Michael's delicious PR comments * Fix off-by-one in tests
2019-12-03 04:28:57 +00:00
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
let _head_timer = metrics::start_timer(&metrics::PERSIST_HEAD);
batch.push(self.persist_head_in_batch());
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
let _fork_choice_timer = metrics::start_timer(&metrics::PERSIST_FORK_CHOICE);
batch.push(self.persist_fork_choice_in_batch());
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
self.store.hot_db.do_atomically(batch)?;
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
Ok(())
}
/// Return a `PersistedBeaconChain` without reference to a `BeaconChain`.
pub fn make_persisted_head(
genesis_block_root: Hash256,
head_tracker: &HeadTracker,
) -> PersistedBeaconChain {
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
PersistedBeaconChain {
_canonical_head_block_root: DUMMY_CANONICAL_HEAD_BLOCK_ROOT,
genesis_block_root,
ssz_head_tracker: head_tracker.to_ssz_container(),
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
}
}
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
/// Return a database operation for writing the beacon chain head to disk.
pub fn persist_head_in_batch(&self) -> KeyValueStoreOp {
Self::persist_head_in_batch_standalone(self.genesis_block_root, &self.head_tracker)
}
pub fn persist_head_in_batch_standalone(
genesis_block_root: Hash256,
head_tracker: &HeadTracker,
) -> KeyValueStoreOp {
Self::make_persisted_head(genesis_block_root, head_tracker)
Fix head tracker concurrency bugs (#1771) ## Issue Addressed Closes #1557 ## Proposed Changes Modify the pruning algorithm so that it mutates the head-tracker _before_ committing the database transaction to disk, and _only if_ all the heads to be removed are still present in the head-tracker (i.e. no concurrent mutations). In the process of writing and testing this I also had to make a few other changes: * Use internal mutability for all `BeaconChainHarness` functions (namely the RNG and the graffiti), in order to enable parallel calls (see testing section below). * Disable logging in harness tests unless the `test_logger` feature is turned on And chose to make some clean-ups: * Delete the `NullMigrator` * Remove type-based configuration for the migrator in favour of runtime config (simpler, less duplicated code) * Use the non-blocking migrator unless the blocking migrator is required. In the store tests we need the blocking migrator because some tests make asserts about the state of the DB after the migration has run. * Rename `validators_keypairs` -> `validator_keypairs` in the `BeaconChainHarness` ## Testing To confirm that the fix worked, I wrote a test using [Hiatus](https://crates.io/crates/hiatus), which can be found here: https://github.com/michaelsproul/lighthouse/tree/hiatus-issue-1557 That test can't be merged because it inserts random breakpoints everywhere, but if you check out that branch you can run the test with: ``` $ cd beacon_node/beacon_chain $ cargo test --release --test parallel_tests --features test_logger ``` It should pass, and the log output should show: ``` WARN Pruning deferred because of a concurrent mutation, message: this is expected only very rarely! ``` ## Additional Info This is a backwards-compatible change with no impact on consensus.
2020-10-19 05:58:39 +00:00
.as_kv_store_op(BEACON_CHAIN_DB_KEY)
}
/// Load fork choice from disk, returning `None` if it isn't found.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub fn load_fork_choice(
store: BeaconStore<T>,
Reset payload statuses when resuming fork choice (#3498) ## Issue Addressed NA ## Proposed Changes This PR is motivated by a recent consensus failure in Geth where it returned `INVALID` for an `VALID` block. Without this PR, the only way to recover is by re-syncing Lighthouse. Whilst ELs "shouldn't have consensus failures", in reality it's something that we can expect from time to time due to the complex nature of Ethereum. Being able to recover easily will help the network recover and EL devs to troubleshoot. The risk introduced with this PR is that genuinely INVALID payloads get a "second chance" at being imported. I believe the DoS risk here is negligible since LH needs to be restarted in order to re-process the payload. Furthermore, there's no reason to think that a well-performing EL will accept a truly invalid payload the second-time-around. ## Additional Info This implementation has the following intricacies: 1. Instead of just resetting *invalid* payloads to optimistic, we'll also reset *valid* payloads. This is an artifact of our existing implementation. 1. We will only reset payload statuses when we detect an invalid payload present in `proto_array` - This helps save us from forgetting that all our blocks are valid in the "best case scenario" where there are no invalid blocks. 1. If we fail to revert the payload statuses we'll log a `CRIT` and just continue with a `proto_array` that *does not* have reverted payload statuses. - The code to revert statuses needs to deal with balances and proposer-boost, so it's a failure point. This is a defensive measure to avoid introducing new show-stopping bugs to LH.
2022-08-29 14:34:41 +00:00
reset_payload_statuses: ResetPayloadStatuses,
count_unrealized_full: CountUnrealizedFull,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
spec: &ChainSpec,
Reset payload statuses when resuming fork choice (#3498) ## Issue Addressed NA ## Proposed Changes This PR is motivated by a recent consensus failure in Geth where it returned `INVALID` for an `VALID` block. Without this PR, the only way to recover is by re-syncing Lighthouse. Whilst ELs "shouldn't have consensus failures", in reality it's something that we can expect from time to time due to the complex nature of Ethereum. Being able to recover easily will help the network recover and EL devs to troubleshoot. The risk introduced with this PR is that genuinely INVALID payloads get a "second chance" at being imported. I believe the DoS risk here is negligible since LH needs to be restarted in order to re-process the payload. Furthermore, there's no reason to think that a well-performing EL will accept a truly invalid payload the second-time-around. ## Additional Info This implementation has the following intricacies: 1. Instead of just resetting *invalid* payloads to optimistic, we'll also reset *valid* payloads. This is an artifact of our existing implementation. 1. We will only reset payload statuses when we detect an invalid payload present in `proto_array` - This helps save us from forgetting that all our blocks are valid in the "best case scenario" where there are no invalid blocks. 1. If we fail to revert the payload statuses we'll log a `CRIT` and just continue with a `proto_array` that *does not* have reverted payload statuses. - The code to revert statuses needs to deal with balances and proposer-boost, so it's a failure point. This is a defensive measure to avoid introducing new show-stopping bugs to LH.
2022-08-29 14:34:41 +00:00
log: &Logger,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
) -> Result<Option<BeaconForkChoice<T>>, Error> {
let persisted_fork_choice =
match store.get_item::<PersistedForkChoice>(&FORK_CHOICE_DB_KEY)? {
Some(fc) => fc,
None => return Ok(None),
};
let fc_store =
BeaconForkChoiceStore::from_persisted(persisted_fork_choice.fork_choice_store, store)?;
Ok(Some(ForkChoice::from_persisted(
persisted_fork_choice.fork_choice,
Reset payload statuses when resuming fork choice (#3498) ## Issue Addressed NA ## Proposed Changes This PR is motivated by a recent consensus failure in Geth where it returned `INVALID` for an `VALID` block. Without this PR, the only way to recover is by re-syncing Lighthouse. Whilst ELs "shouldn't have consensus failures", in reality it's something that we can expect from time to time due to the complex nature of Ethereum. Being able to recover easily will help the network recover and EL devs to troubleshoot. The risk introduced with this PR is that genuinely INVALID payloads get a "second chance" at being imported. I believe the DoS risk here is negligible since LH needs to be restarted in order to re-process the payload. Furthermore, there's no reason to think that a well-performing EL will accept a truly invalid payload the second-time-around. ## Additional Info This implementation has the following intricacies: 1. Instead of just resetting *invalid* payloads to optimistic, we'll also reset *valid* payloads. This is an artifact of our existing implementation. 1. We will only reset payload statuses when we detect an invalid payload present in `proto_array` - This helps save us from forgetting that all our blocks are valid in the "best case scenario" where there are no invalid blocks. 1. If we fail to revert the payload statuses we'll log a `CRIT` and just continue with a `proto_array` that *does not* have reverted payload statuses. - The code to revert statuses needs to deal with balances and proposer-boost, so it's a failure point. This is a defensive measure to avoid introducing new show-stopping bugs to LH.
2022-08-29 14:34:41 +00:00
reset_payload_statuses,
fc_store,
count_unrealized_full,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
spec,
Reset payload statuses when resuming fork choice (#3498) ## Issue Addressed NA ## Proposed Changes This PR is motivated by a recent consensus failure in Geth where it returned `INVALID` for an `VALID` block. Without this PR, the only way to recover is by re-syncing Lighthouse. Whilst ELs "shouldn't have consensus failures", in reality it's something that we can expect from time to time due to the complex nature of Ethereum. Being able to recover easily will help the network recover and EL devs to troubleshoot. The risk introduced with this PR is that genuinely INVALID payloads get a "second chance" at being imported. I believe the DoS risk here is negligible since LH needs to be restarted in order to re-process the payload. Furthermore, there's no reason to think that a well-performing EL will accept a truly invalid payload the second-time-around. ## Additional Info This implementation has the following intricacies: 1. Instead of just resetting *invalid* payloads to optimistic, we'll also reset *valid* payloads. This is an artifact of our existing implementation. 1. We will only reset payload statuses when we detect an invalid payload present in `proto_array` - This helps save us from forgetting that all our blocks are valid in the "best case scenario" where there are no invalid blocks. 1. If we fail to revert the payload statuses we'll log a `CRIT` and just continue with a `proto_array` that *does not* have reverted payload statuses. - The code to revert statuses needs to deal with balances and proposer-boost, so it's a failure point. This is a defensive measure to avoid introducing new show-stopping bugs to LH.
2022-08-29 14:34:41 +00:00
log,
)?))
}
/// Persists `self.op_pool` to disk.
///
/// ## Notes
///
/// This operation is typically slow and causes a lot of allocations. It should be used
/// sparingly.
pub fn persist_op_pool(&self) -> Result<(), Error> {
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
let _timer = metrics::start_timer(&metrics::PERSIST_OP_POOL);
self.store.put_item(
&OP_POOL_DB_KEY,
&PersistedOperationPool::from_operation_pool(&self.op_pool),
)?;
Ok(())
}
/// Persists `self.eth1_chain` and its caches to disk.
pub fn persist_eth1_cache(&self) -> Result<(), Error> {
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
let _timer = metrics::start_timer(&metrics::PERSIST_OP_POOL);
if let Some(eth1_chain) = self.eth1_chain.as_ref() {
self.store
.put_item(&ETH1_CACHE_DB_KEY, &eth1_chain.as_ssz_container())?;
}
Ok(())
}
2019-08-29 04:26:30 +00:00
/// Returns the slot _right now_ according to `self.slot_clock`. Returns `Err` if the slot is
/// unavailable.
///
/// The slot might be unavailable due to an error with the system clock, or if the present time
/// is before genesis (i.e., a negative slot).
2019-08-29 04:26:30 +00:00
pub fn slot(&self) -> Result<Slot, Error> {
self.slot_clock.now().ok_or(Error::UnableToReadSlot)
}
/// Returns the epoch _right now_ according to `self.slot_clock`. Returns `Err` if the epoch is
/// unavailable.
///
/// The epoch might be unavailable due to an error with the system clock, or if the present time
/// is before genesis (i.e., a negative epoch).
pub fn epoch(&self) -> Result<Epoch, Error> {
self.slot()
.map(|slot| slot.epoch(T::EthSpec::slots_per_epoch()))
}
/// Iterates across all `(block_root, slot)` pairs from `start_slot`
/// to the head of the chain (inclusive).
2019-06-05 03:33:32 +00:00
///
/// ## Notes
2019-06-05 03:33:32 +00:00
///
/// - `slot` always increases by `1`.
/// - Skipped slots contain the root of the closest prior
/// non-skipped slot (identical to the way they are stored in `state.block_roots`).
/// - Iterator returns `(Hash256, Slot)`.
///
/// Will return a `BlockOutOfRange` error if the requested start slot is before the period of
/// history for which we have blocks stored. See `get_oldest_block_slot`.
pub fn forwards_iter_block_roots(
&self,
start_slot: Slot,
) -> Result<impl Iterator<Item = Result<(Hash256, Slot), Error>> + '_, Error> {
let oldest_block_slot = self.store.get_oldest_block_slot();
if start_slot < oldest_block_slot {
return Err(Error::HistoricalBlockError(
HistoricalBlockError::BlockOutOfRange {
slot: start_slot,
oldest_block_slot,
},
));
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let local_head = self.head_snapshot();
let iter = self.store.forwards_block_roots_iterator(
start_slot,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
local_head.beacon_state.clone_with(CloneConfig::none()),
local_head.beacon_block_root,
&self.spec,
)?;
Ok(iter.map(|result| result.map_err(Into::into)))
}
/// Even more efficient variant of `forwards_iter_block_roots` that will avoid cloning the head
/// state if it isn't required for the requested range of blocks.
/// The range [start_slot, end_slot] is inclusive (ie `start_slot <= end_slot`)
pub fn forwards_iter_block_roots_until(
&self,
start_slot: Slot,
end_slot: Slot,
) -> Result<impl Iterator<Item = Result<(Hash256, Slot), Error>> + '_, Error> {
let oldest_block_slot = self.store.get_oldest_block_slot();
if start_slot < oldest_block_slot {
return Err(Error::HistoricalBlockError(
HistoricalBlockError::BlockOutOfRange {
slot: start_slot,
oldest_block_slot,
},
));
}
self.with_head(move |head| {
let iter = self.store.forwards_block_roots_iterator_until(
start_slot,
end_slot,
|| {
(
head.beacon_state.clone_with_only_committee_caches(),
head.beacon_block_root,
)
},
&self.spec,
)?;
Ok(iter
.map(|result| result.map_err(Into::into))
.take_while(move |result| {
result.as_ref().map_or(true, |(_, slot)| *slot <= end_slot)
}))
})
}
/// Traverse backwards from `block_root` to find the block roots of its ancestors.
///
/// ## Notes
///
/// - `slot` always decreases by `1`.
/// - Skipped slots contain the root of the closest prior
/// non-skipped slot (identical to the way they are stored in `state.block_roots`) .
/// - Iterator returns `(Hash256, Slot)`.
/// - The provided `block_root` is included as the first item in the iterator.
pub fn rev_iter_block_roots_from(
&self,
block_root: Hash256,
) -> Result<impl Iterator<Item = Result<(Hash256, Slot), Error>> + '_, Error> {
let block = self
Separate execution payloads in the DB (#3157) ## Proposed Changes Reduce post-merge disk usage by not storing finalized execution payloads in Lighthouse's database. :warning: **This is achieved in a backwards-incompatible way for networks that have already merged** :warning:. Kiln users and shadow fork enjoyers will be unable to downgrade after running the code from this PR. The upgrade migration may take several minutes to run, and can't be aborted after it begins. The main changes are: - New column in the database called `ExecPayload`, keyed by beacon block root. - The `BeaconBlock` column now stores blinded blocks only. - Lots of places that previously used full blocks now use blinded blocks, e.g. analytics APIs, block replay in the DB, etc. - On finalization: - `prune_abanonded_forks` deletes non-canonical payloads whilst deleting non-canonical blocks. - `migrate_db` deletes finalized canonical payloads whilst deleting finalized states. - Conversions between blinded and full blocks are implemented in a compositional way, duplicating some work from Sean's PR #3134. - The execution layer has a new `get_payload_by_block_hash` method that reconstructs a payload using the EE's `eth_getBlockByHash` call. - I've tested manually that it works on Kiln, using Geth and Nethermind. - This isn't necessarily the most efficient method, and new engine APIs are being discussed to improve this: https://github.com/ethereum/execution-apis/pull/146. - We're depending on the `ethers` master branch, due to lots of recent changes. We're also using a workaround for https://github.com/gakonst/ethers-rs/issues/1134. - Payload reconstruction is used in the HTTP API via `BeaconChain::get_block`, which is now `async`. Due to the `async` fn, the `blocking_json` wrapper has been removed. - Payload reconstruction is used in network RPC to serve blocks-by-{root,range} responses. Here the `async` adjustment is messier, although I think I've managed to come up with a reasonable compromise: the handlers take the `SendOnDrop` by value so that they can drop it on _task completion_ (after the `fn` returns). Still, this is introducing disk reads onto core executor threads, which may have a negative performance impact (thoughts appreciated). ## Additional Info - [x] For performance it would be great to remove the cloning of full blocks when converting them to blinded blocks to write to disk. I'm going to experiment with a `put_block` API that takes the block by value, breaks it into a blinded block and a payload, stores the blinded block, and then re-assembles the full block for the caller. - [x] We should measure the latency of blocks-by-root and blocks-by-range responses. - [x] We should add integration tests that stress the payload reconstruction (basic tests done, issue for more extensive tests: https://github.com/sigp/lighthouse/issues/3159) - [x] We should (manually) test the schema v9 migration from several prior versions, particularly as blocks have changed on disk and some migrations rely on being able to load blocks. Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-05-12 00:42:17 +00:00
.get_blinded_block(&block_root)?
.ok_or(Error::MissingBeaconBlock(block_root))?;
let state = self
Update to Spec v0.10 (#817) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Disable/delete two outdated tests * Bump eth1 default vote warn to error * Delete outdated eth1 test Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-02-10 23:19:36 +00:00
.get_state(&block.state_root(), Some(block.slot()))?
.ok_or_else(|| Error::MissingBeaconState(block.state_root()))?;
let iter = BlockRootsIterator::owned(&self.store, state);
Ok(std::iter::once(Ok((block_root, block.slot())))
.chain(iter)
.map(|result| result.map_err(|e| e.into())))
}
/// Iterates backwards across all `(state_root, slot)` pairs starting from
/// an arbitrary `BeaconState` to the earliest reachable ancestor (may or may not be genesis).
///
/// ## Notes
///
/// - `slot` always decreases by `1`.
/// - Iterator returns `(Hash256, Slot)`.
/// - As this iterator starts at the `head` of the chain (viz., the best block), the first slot
/// returned may be earlier than the wall-clock slot.
pub fn rev_iter_state_roots_from<'a>(
&'a self,
state_root: Hash256,
state: &'a BeaconState<T::EthSpec>,
) -> impl Iterator<Item = Result<(Hash256, Slot), Error>> + 'a {
std::iter::once(Ok((state_root, state.slot())))
.chain(StateRootsIterator::new(&self.store, state))
.map(|result| result.map_err(Into::into))
}
/// Iterates across all `(state_root, slot)` pairs from `start_slot`
/// to the head of the chain (inclusive).
///
/// ## Notes
///
/// - `slot` always increases by `1`.
/// - Iterator returns `(Hash256, Slot)`.
pub fn forwards_iter_state_roots(
&self,
start_slot: Slot,
) -> Result<impl Iterator<Item = Result<(Hash256, Slot), Error>> + '_, Error> {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let local_head = self.head_snapshot();
let iter = self.store.forwards_state_roots_iterator(
start_slot,
local_head.beacon_state_root(),
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
local_head.beacon_state.clone_with(CloneConfig::none()),
&self.spec,
)?;
Ok(iter.map(|result| result.map_err(Into::into)))
}
/// Super-efficient forwards state roots iterator that avoids cloning the head if the state
/// roots lie entirely within the freezer database.
///
/// The iterator returned will include roots for `start_slot..=end_slot`, i.e. it
/// is endpoint inclusive.
pub fn forwards_iter_state_roots_until(
&self,
start_slot: Slot,
end_slot: Slot,
) -> Result<impl Iterator<Item = Result<(Hash256, Slot), Error>> + '_, Error> {
self.with_head(move |head| {
let iter = self.store.forwards_state_roots_iterator_until(
start_slot,
end_slot,
|| {
(
head.beacon_state.clone_with_only_committee_caches(),
head.beacon_state_root(),
)
},
&self.spec,
)?;
Ok(iter
.map(|result| result.map_err(Into::into))
.take_while(move |result| {
result.as_ref().map_or(true, |(_, slot)| *slot <= end_slot)
}))
})
}
Validator client refactor (#618) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Address Michael's comments
2019-11-25 04:48:24 +00:00
/// Returns the block at the given slot, if any. Only returns blocks in the canonical chain.
///
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
/// Use the `skips` parameter to define the behaviour when `request_slot` is a skipped slot.
///
Validator client refactor (#618) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Address Michael's comments
2019-11-25 04:48:24 +00:00
/// ## Errors
///
/// May return a database error.
Update to Spec v0.10 (#817) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Disable/delete two outdated tests * Bump eth1 default vote warn to error * Delete outdated eth1 test Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-02-10 23:19:36 +00:00
pub fn block_at_slot(
&self,
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
request_slot: Slot,
skips: WhenSlotSkipped,
Separate execution payloads in the DB (#3157) ## Proposed Changes Reduce post-merge disk usage by not storing finalized execution payloads in Lighthouse's database. :warning: **This is achieved in a backwards-incompatible way for networks that have already merged** :warning:. Kiln users and shadow fork enjoyers will be unable to downgrade after running the code from this PR. The upgrade migration may take several minutes to run, and can't be aborted after it begins. The main changes are: - New column in the database called `ExecPayload`, keyed by beacon block root. - The `BeaconBlock` column now stores blinded blocks only. - Lots of places that previously used full blocks now use blinded blocks, e.g. analytics APIs, block replay in the DB, etc. - On finalization: - `prune_abanonded_forks` deletes non-canonical payloads whilst deleting non-canonical blocks. - `migrate_db` deletes finalized canonical payloads whilst deleting finalized states. - Conversions between blinded and full blocks are implemented in a compositional way, duplicating some work from Sean's PR #3134. - The execution layer has a new `get_payload_by_block_hash` method that reconstructs a payload using the EE's `eth_getBlockByHash` call. - I've tested manually that it works on Kiln, using Geth and Nethermind. - This isn't necessarily the most efficient method, and new engine APIs are being discussed to improve this: https://github.com/ethereum/execution-apis/pull/146. - We're depending on the `ethers` master branch, due to lots of recent changes. We're also using a workaround for https://github.com/gakonst/ethers-rs/issues/1134. - Payload reconstruction is used in the HTTP API via `BeaconChain::get_block`, which is now `async`. Due to the `async` fn, the `blocking_json` wrapper has been removed. - Payload reconstruction is used in network RPC to serve blocks-by-{root,range} responses. Here the `async` adjustment is messier, although I think I've managed to come up with a reasonable compromise: the handlers take the `SendOnDrop` by value so that they can drop it on _task completion_ (after the `fn` returns). Still, this is introducing disk reads onto core executor threads, which may have a negative performance impact (thoughts appreciated). ## Additional Info - [x] For performance it would be great to remove the cloning of full blocks when converting them to blinded blocks to write to disk. I'm going to experiment with a `put_block` API that takes the block by value, breaks it into a blinded block and a payload, stores the blinded block, and then re-assembles the full block for the caller. - [x] We should measure the latency of blocks-by-root and blocks-by-range responses. - [x] We should add integration tests that stress the payload reconstruction (basic tests done, issue for more extensive tests: https://github.com/sigp/lighthouse/issues/3159) - [x] We should (manually) test the schema v9 migration from several prior versions, particularly as blocks have changed on disk and some migrations rely on being able to load blocks. Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-05-12 00:42:17 +00:00
) -> Result<Option<SignedBlindedBeaconBlock<T::EthSpec>>, Error> {
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
let root = self.block_root_at_slot(request_slot, skips)?;
Validator client refactor (#618) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Address Michael's comments
2019-11-25 04:48:24 +00:00
if let Some(block_root) = root {
Separate execution payloads in the DB (#3157) ## Proposed Changes Reduce post-merge disk usage by not storing finalized execution payloads in Lighthouse's database. :warning: **This is achieved in a backwards-incompatible way for networks that have already merged** :warning:. Kiln users and shadow fork enjoyers will be unable to downgrade after running the code from this PR. The upgrade migration may take several minutes to run, and can't be aborted after it begins. The main changes are: - New column in the database called `ExecPayload`, keyed by beacon block root. - The `BeaconBlock` column now stores blinded blocks only. - Lots of places that previously used full blocks now use blinded blocks, e.g. analytics APIs, block replay in the DB, etc. - On finalization: - `prune_abanonded_forks` deletes non-canonical payloads whilst deleting non-canonical blocks. - `migrate_db` deletes finalized canonical payloads whilst deleting finalized states. - Conversions between blinded and full blocks are implemented in a compositional way, duplicating some work from Sean's PR #3134. - The execution layer has a new `get_payload_by_block_hash` method that reconstructs a payload using the EE's `eth_getBlockByHash` call. - I've tested manually that it works on Kiln, using Geth and Nethermind. - This isn't necessarily the most efficient method, and new engine APIs are being discussed to improve this: https://github.com/ethereum/execution-apis/pull/146. - We're depending on the `ethers` master branch, due to lots of recent changes. We're also using a workaround for https://github.com/gakonst/ethers-rs/issues/1134. - Payload reconstruction is used in the HTTP API via `BeaconChain::get_block`, which is now `async`. Due to the `async` fn, the `blocking_json` wrapper has been removed. - Payload reconstruction is used in network RPC to serve blocks-by-{root,range} responses. Here the `async` adjustment is messier, although I think I've managed to come up with a reasonable compromise: the handlers take the `SendOnDrop` by value so that they can drop it on _task completion_ (after the `fn` returns). Still, this is introducing disk reads onto core executor threads, which may have a negative performance impact (thoughts appreciated). ## Additional Info - [x] For performance it would be great to remove the cloning of full blocks when converting them to blinded blocks to write to disk. I'm going to experiment with a `put_block` API that takes the block by value, breaks it into a blinded block and a payload, stores the blinded block, and then re-assembles the full block for the caller. - [x] We should measure the latency of blocks-by-root and blocks-by-range responses. - [x] We should add integration tests that stress the payload reconstruction (basic tests done, issue for more extensive tests: https://github.com/sigp/lighthouse/issues/3159) - [x] We should (manually) test the schema v9 migration from several prior versions, particularly as blocks have changed on disk and some migrations rely on being able to load blocks. Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-05-12 00:42:17 +00:00
Ok(self.store.get_blinded_block(&block_root)?)
Validator client refactor (#618) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Address Michael's comments
2019-11-25 04:48:24 +00:00
} else {
Ok(None)
}
}
/// Returns the state root at the given slot, if any. Only returns state roots in the canonical chain.
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
///
/// ## Errors
///
/// May return a database error.
pub fn state_root_at_slot(&self, request_slot: Slot) -> Result<Option<Hash256>, Error> {
if request_slot > self.slot()? {
return Ok(None);
} else if request_slot == self.spec.genesis_slot {
return Ok(Some(self.genesis_state_root));
}
// Check limits w.r.t historic state bounds.
let (historic_lower_limit, historic_upper_limit) = self.store.get_historic_state_limits();
if request_slot > historic_lower_limit && request_slot < historic_upper_limit {
return Ok(None);
}
// Try an optimized path of reading the root directly from the head state.
let fast_lookup: Option<Hash256> = self.with_head(|head| {
if head.beacon_block.slot() <= request_slot {
// Return the head state root if all slots between the request and the head are skipped.
Ok(Some(head.beacon_state_root()))
} else if let Ok(root) = head.beacon_state.get_state_root(request_slot) {
// Return the root if it's easily accessible from the head state.
Ok(Some(*root))
} else {
// Fast lookup is not possible.
Ok::<_, Error>(None)
}
})?;
if let Some(root) = fast_lookup {
return Ok(Some(root));
}
process_results(
self.forwards_iter_state_roots_until(request_slot, request_slot)?,
|mut iter| {
if let Some((root, slot)) = iter.next() {
if slot == request_slot {
Ok(Some(root))
} else {
// Sanity check.
Err(Error::InconsistentForwardsIter { request_slot, slot })
}
} else {
Ok(None)
}
},
)?
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
}
/// Returns the block root at the given slot, if any. Only returns roots in the canonical chain.
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
///
/// ## Notes
///
/// - Use the `skips` parameter to define the behaviour when `request_slot` is a skipped slot.
/// - Returns `Ok(None)` for any slot higher than the current wall-clock slot, or less than
/// the oldest known block slot.
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
pub fn block_root_at_slot(
&self,
request_slot: Slot,
skips: WhenSlotSkipped,
) -> Result<Option<Hash256>, Error> {
match skips {
WhenSlotSkipped::None => self.block_root_at_slot_skips_none(request_slot),
WhenSlotSkipped::Prev => self.block_root_at_slot_skips_prev(request_slot),
}
.or_else(|e| match e {
Error::HistoricalBlockError(_) => Ok(None),
e => Err(e),
})
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
}
/// Returns the block root at the given slot, if any. Only returns roots in the canonical chain.
///
/// ## Notes
///
/// - Returns `Ok(None)` if the given `Slot` was skipped.
/// - Returns `Ok(None)` for any slot higher than the current wall-clock slot.
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
///
/// ## Errors
///
/// May return a database error.
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
fn block_root_at_slot_skips_none(&self, request_slot: Slot) -> Result<Option<Hash256>, Error> {
if request_slot > self.slot()? {
return Ok(None);
} else if request_slot == self.spec.genesis_slot {
return Ok(Some(self.genesis_block_root));
}
let prev_slot = request_slot.saturating_sub(1_u64);
// Try an optimized path of reading the root directly from the head state.
let fast_lookup: Option<Option<Hash256>> = self.with_head(|head| {
let state = &head.beacon_state;
// Try find the root for the `request_slot`.
let request_root_opt = match state.slot().cmp(&request_slot) {
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
// It's always a skip slot if the head is less than the request slot, return early.
Ordering::Less => return Ok(Some(None)),
// The request slot is the head slot.
Ordering::Equal => Some(head.beacon_block_root),
// Try find the request slot in the state.
Ordering::Greater => state.get_block_root(request_slot).ok().copied(),
};
if let Some(request_root) = request_root_opt {
if let Ok(prev_root) = state.get_block_root(prev_slot) {
return Ok(Some((*prev_root != request_root).then_some(request_root)));
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
}
}
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
// Fast lookup is not possible.
Ok::<_, Error>(None)
})?;
if let Some(root_opt) = fast_lookup {
return Ok(root_opt);
}
if let Some(((prev_root, _), (curr_root, curr_slot))) = process_results(
self.forwards_iter_block_roots_until(prev_slot, request_slot)?,
|iter| iter.tuple_windows().next(),
)? {
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
// Sanity check.
if curr_slot != request_slot {
return Err(Error::InconsistentForwardsIter {
request_slot,
slot: curr_slot,
});
}
Ok((curr_root != prev_root).then_some(curr_root))
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
} else {
Ok(None)
}
}
/// Returns the block root at the given slot, if any. Only returns roots in the canonical chain.
///
/// ## Notes
///
/// - Returns the root at the previous non-skipped slot if the given `Slot` was skipped.
/// - Returns `Ok(None)` for any slot higher than the current wall-clock slot.
///
/// ## Errors
///
/// May return a database error.
fn block_root_at_slot_skips_prev(&self, request_slot: Slot) -> Result<Option<Hash256>, Error> {
if request_slot > self.slot()? {
return Ok(None);
} else if request_slot == self.spec.genesis_slot {
return Ok(Some(self.genesis_block_root));
}
// Try an optimized path of reading the root directly from the head state.
let fast_lookup: Option<Hash256> = self.with_head(|head| {
if head.beacon_block.slot() <= request_slot {
// Return the head root if all slots between the request and the head are skipped.
Ok(Some(head.beacon_block_root))
} else if let Ok(root) = head.beacon_state.get_block_root(request_slot) {
// Return the root if it's easily accessible from the head state.
Ok(Some(*root))
} else {
// Fast lookup is not possible.
Ok::<_, Error>(None)
}
})?;
if let Some(root) = fast_lookup {
return Ok(Some(root));
}
process_results(
self.forwards_iter_block_roots_until(request_slot, request_slot)?,
|mut iter| {
if let Some((root, slot)) = iter.next() {
if slot == request_slot {
Ok(Some(root))
} else {
// Sanity check.
Err(Error::InconsistentForwardsIter { request_slot, slot })
}
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
} else {
Ok(None)
Use the forwards iterator more often (#2376) ## Issue Addressed NA ## Primary Change When investigating memory usage, I noticed that retrieving a block from an early slot (e.g., slot 900) would cause a sharp increase in the memory footprint (from 400mb to 800mb+) which seemed to be ever-lasting. After some investigation, I found that the reverse iteration from the head back to that slot was the likely culprit. To counter this, I've switched the `BeaconChain::block_root_at_slot` to use the forwards iterator, instead of the reverse one. I also noticed that the networking stack is using `BeaconChain::root_at_slot` to check if a peer is relevant (`check_peer_relevance`). Perhaps the steep, seemingly-random-but-consistent increases in memory usage are caused by the use of this function. Using the forwards iterator with the HTTP API alleviated the sharp increases in memory usage. It also made the response much faster (before it felt like to took 1-2s, now it feels instant). ## Additional Changes In the process I also noticed that we have two functions for getting block roots: - `BeaconChain::block_root_at_slot`: returns `None` for a skip slot. - `BeaconChain::root_at_slot`: returns the previous root for a skip slot. I unified these two functions into `block_root_at_slot` and added the `WhenSlotSkipped` enum. Now, the caller must be explicit about the skip-slot behaviour when requesting a root. Additionally, I replaced `vec![]` with `Vec::with_capacity` in `store::chunked_vector::range_query`. I stumbled across this whilst debugging and made this modification to see what effect it would have (not much). It seems like a decent change to keep around, but I'm not concerned either way. Also, `BeaconChain::get_ancestor_block_root` is unused, so I got rid of it :wastebasket:. ## Additional Info I haven't also done the same for state roots here. Whilst it's possible and a good idea, it's more work since the fwds iterators are presently block-roots-specific. Whilst there's a few places a reverse iteration of state roots could be triggered (e.g., attestation production, HTTP API), they're no where near as common as the `check_peer_relevance` call. As such, I think we should get this PR merged first, then come back for the state root iters. I made an issue here https://github.com/sigp/lighthouse/issues/2377.
2021-05-31 04:18:20 +00:00
}
},
)?
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
}
Add early attester cache (#2872) ## Issue Addressed NA ## Proposed Changes Introduces a cache to attestation to produce atop blocks which will become the head, but are not fully imported (e.g., not inserted into the database). Whilst attesting to a block before it's imported is rather easy, if we're going to produce that attestation then we also need to be able to: 1. Verify that attestation. 1. Respond to RPC requests for the `beacon_block_root`. Attestation verification (1) is *partially* covered. Since we prime the shuffling cache before we insert the block into the early attester cache, we should be fine for all typical use-cases. However, it is possible that the cache is washed out before we've managed to insert the state into the database and then attestation verification will fail with a "missing beacon state"-type error. Providing the block via RPC (2) is also partially covered, since we'll check the database *and* the early attester cache when responding a blocks-by-root request. However, we'll still omit the block from blocks-by-range requests (until the block lands in the DB). I *think* this is fine, since there's no guarantee that we return all blocks for those responses. Another important consideration is whether or not the *parent* of the early attester block is available in the databse. If it were not, we might fail to respond to blocks-by-root request that are iterating backwards to collect a chain of blocks. I argue that *we will always have the parent of the early attester block in the database.* This is because we are holding the fork-choice write-lock when inserting the block into the early attester cache and we do not drop that until the block is in the database.
2022-01-11 01:35:55 +00:00
/// Returns the block at the given root, if any.
///
/// Will also check the early attester cache for the block. Because of this, there's no
/// guarantee that a block returned from this function has a `BeaconState` available in
/// `self.store`. The expected use for this function is *only* for returning blocks requested
/// from P2P peers.
///
/// ## Errors
///
/// May return a database error.
Separate execution payloads in the DB (#3157) ## Proposed Changes Reduce post-merge disk usage by not storing finalized execution payloads in Lighthouse's database. :warning: **This is achieved in a backwards-incompatible way for networks that have already merged** :warning:. Kiln users and shadow fork enjoyers will be unable to downgrade after running the code from this PR. The upgrade migration may take several minutes to run, and can't be aborted after it begins. The main changes are: - New column in the database called `ExecPayload`, keyed by beacon block root. - The `BeaconBlock` column now stores blinded blocks only. - Lots of places that previously used full blocks now use blinded blocks, e.g. analytics APIs, block replay in the DB, etc. - On finalization: - `prune_abanonded_forks` deletes non-canonical payloads whilst deleting non-canonical blocks. - `migrate_db` deletes finalized canonical payloads whilst deleting finalized states. - Conversions between blinded and full blocks are implemented in a compositional way, duplicating some work from Sean's PR #3134. - The execution layer has a new `get_payload_by_block_hash` method that reconstructs a payload using the EE's `eth_getBlockByHash` call. - I've tested manually that it works on Kiln, using Geth and Nethermind. - This isn't necessarily the most efficient method, and new engine APIs are being discussed to improve this: https://github.com/ethereum/execution-apis/pull/146. - We're depending on the `ethers` master branch, due to lots of recent changes. We're also using a workaround for https://github.com/gakonst/ethers-rs/issues/1134. - Payload reconstruction is used in the HTTP API via `BeaconChain::get_block`, which is now `async`. Due to the `async` fn, the `blocking_json` wrapper has been removed. - Payload reconstruction is used in network RPC to serve blocks-by-{root,range} responses. Here the `async` adjustment is messier, although I think I've managed to come up with a reasonable compromise: the handlers take the `SendOnDrop` by value so that they can drop it on _task completion_ (after the `fn` returns). Still, this is introducing disk reads onto core executor threads, which may have a negative performance impact (thoughts appreciated). ## Additional Info - [x] For performance it would be great to remove the cloning of full blocks when converting them to blinded blocks to write to disk. I'm going to experiment with a `put_block` API that takes the block by value, breaks it into a blinded block and a payload, stores the blinded block, and then re-assembles the full block for the caller. - [x] We should measure the latency of blocks-by-root and blocks-by-range responses. - [x] We should add integration tests that stress the payload reconstruction (basic tests done, issue for more extensive tests: https://github.com/sigp/lighthouse/issues/3159) - [x] We should (manually) test the schema v9 migration from several prior versions, particularly as blocks have changed on disk and some migrations rely on being able to load blocks. Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-05-12 00:42:17 +00:00
pub async fn get_block_checking_early_attester_cache(
Add early attester cache (#2872) ## Issue Addressed NA ## Proposed Changes Introduces a cache to attestation to produce atop blocks which will become the head, but are not fully imported (e.g., not inserted into the database). Whilst attesting to a block before it's imported is rather easy, if we're going to produce that attestation then we also need to be able to: 1. Verify that attestation. 1. Respond to RPC requests for the `beacon_block_root`. Attestation verification (1) is *partially* covered. Since we prime the shuffling cache before we insert the block into the early attester cache, we should be fine for all typical use-cases. However, it is possible that the cache is washed out before we've managed to insert the state into the database and then attestation verification will fail with a "missing beacon state"-type error. Providing the block via RPC (2) is also partially covered, since we'll check the database *and* the early attester cache when responding a blocks-by-root request. However, we'll still omit the block from blocks-by-range requests (until the block lands in the DB). I *think* this is fine, since there's no guarantee that we return all blocks for those responses. Another important consideration is whether or not the *parent* of the early attester block is available in the databse. If it were not, we might fail to respond to blocks-by-root request that are iterating backwards to collect a chain of blocks. I argue that *we will always have the parent of the early attester block in the database.* This is because we are holding the fork-choice write-lock when inserting the block into the early attester cache and we do not drop that until the block is in the database.
2022-01-11 01:35:55 +00:00
&self,
block_root: &Hash256,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
) -> Result<Option<Arc<SignedBeaconBlock<T::EthSpec>>>, Error> {
Separate execution payloads in the DB (#3157) ## Proposed Changes Reduce post-merge disk usage by not storing finalized execution payloads in Lighthouse's database. :warning: **This is achieved in a backwards-incompatible way for networks that have already merged** :warning:. Kiln users and shadow fork enjoyers will be unable to downgrade after running the code from this PR. The upgrade migration may take several minutes to run, and can't be aborted after it begins. The main changes are: - New column in the database called `ExecPayload`, keyed by beacon block root. - The `BeaconBlock` column now stores blinded blocks only. - Lots of places that previously used full blocks now use blinded blocks, e.g. analytics APIs, block replay in the DB, etc. - On finalization: - `prune_abanonded_forks` deletes non-canonical payloads whilst deleting non-canonical blocks. - `migrate_db` deletes finalized canonical payloads whilst deleting finalized states. - Conversions between blinded and full blocks are implemented in a compositional way, duplicating some work from Sean's PR #3134. - The execution layer has a new `get_payload_by_block_hash` method that reconstructs a payload using the EE's `eth_getBlockByHash` call. - I've tested manually that it works on Kiln, using Geth and Nethermind. - This isn't necessarily the most efficient method, and new engine APIs are being discussed to improve this: https://github.com/ethereum/execution-apis/pull/146. - We're depending on the `ethers` master branch, due to lots of recent changes. We're also using a workaround for https://github.com/gakonst/ethers-rs/issues/1134. - Payload reconstruction is used in the HTTP API via `BeaconChain::get_block`, which is now `async`. Due to the `async` fn, the `blocking_json` wrapper has been removed. - Payload reconstruction is used in network RPC to serve blocks-by-{root,range} responses. Here the `async` adjustment is messier, although I think I've managed to come up with a reasonable compromise: the handlers take the `SendOnDrop` by value so that they can drop it on _task completion_ (after the `fn` returns). Still, this is introducing disk reads onto core executor threads, which may have a negative performance impact (thoughts appreciated). ## Additional Info - [x] For performance it would be great to remove the cloning of full blocks when converting them to blinded blocks to write to disk. I'm going to experiment with a `put_block` API that takes the block by value, breaks it into a blinded block and a payload, stores the blinded block, and then re-assembles the full block for the caller. - [x] We should measure the latency of blocks-by-root and blocks-by-range responses. - [x] We should add integration tests that stress the payload reconstruction (basic tests done, issue for more extensive tests: https://github.com/sigp/lighthouse/issues/3159) - [x] We should (manually) test the schema v9 migration from several prior versions, particularly as blocks have changed on disk and some migrations rely on being able to load blocks. Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-05-12 00:42:17 +00:00
if let Some(block) = self.early_attester_cache.get_block(*block_root) {
return Ok(Some(block));
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
Ok(self.get_block(block_root).await?.map(Arc::new))
Add early attester cache (#2872) ## Issue Addressed NA ## Proposed Changes Introduces a cache to attestation to produce atop blocks which will become the head, but are not fully imported (e.g., not inserted into the database). Whilst attesting to a block before it's imported is rather easy, if we're going to produce that attestation then we also need to be able to: 1. Verify that attestation. 1. Respond to RPC requests for the `beacon_block_root`. Attestation verification (1) is *partially* covered. Since we prime the shuffling cache before we insert the block into the early attester cache, we should be fine for all typical use-cases. However, it is possible that the cache is washed out before we've managed to insert the state into the database and then attestation verification will fail with a "missing beacon state"-type error. Providing the block via RPC (2) is also partially covered, since we'll check the database *and* the early attester cache when responding a blocks-by-root request. However, we'll still omit the block from blocks-by-range requests (until the block lands in the DB). I *think* this is fine, since there's no guarantee that we return all blocks for those responses. Another important consideration is whether or not the *parent* of the early attester block is available in the databse. If it were not, we might fail to respond to blocks-by-root request that are iterating backwards to collect a chain of blocks. I argue that *we will always have the parent of the early attester block in the database.* This is because we are holding the fork-choice write-lock when inserting the block into the early attester cache and we do not drop that until the block is in the database.
2022-01-11 01:35:55 +00:00
}
2019-03-23 02:23:44 +00:00
/// Returns the block at the given root, if any.
///
/// ## Errors
///
/// May return a database error.
Separate execution payloads in the DB (#3157) ## Proposed Changes Reduce post-merge disk usage by not storing finalized execution payloads in Lighthouse's database. :warning: **This is achieved in a backwards-incompatible way for networks that have already merged** :warning:. Kiln users and shadow fork enjoyers will be unable to downgrade after running the code from this PR. The upgrade migration may take several minutes to run, and can't be aborted after it begins. The main changes are: - New column in the database called `ExecPayload`, keyed by beacon block root. - The `BeaconBlock` column now stores blinded blocks only. - Lots of places that previously used full blocks now use blinded blocks, e.g. analytics APIs, block replay in the DB, etc. - On finalization: - `prune_abanonded_forks` deletes non-canonical payloads whilst deleting non-canonical blocks. - `migrate_db` deletes finalized canonical payloads whilst deleting finalized states. - Conversions between blinded and full blocks are implemented in a compositional way, duplicating some work from Sean's PR #3134. - The execution layer has a new `get_payload_by_block_hash` method that reconstructs a payload using the EE's `eth_getBlockByHash` call. - I've tested manually that it works on Kiln, using Geth and Nethermind. - This isn't necessarily the most efficient method, and new engine APIs are being discussed to improve this: https://github.com/ethereum/execution-apis/pull/146. - We're depending on the `ethers` master branch, due to lots of recent changes. We're also using a workaround for https://github.com/gakonst/ethers-rs/issues/1134. - Payload reconstruction is used in the HTTP API via `BeaconChain::get_block`, which is now `async`. Due to the `async` fn, the `blocking_json` wrapper has been removed. - Payload reconstruction is used in network RPC to serve blocks-by-{root,range} responses. Here the `async` adjustment is messier, although I think I've managed to come up with a reasonable compromise: the handlers take the `SendOnDrop` by value so that they can drop it on _task completion_ (after the `fn` returns). Still, this is introducing disk reads onto core executor threads, which may have a negative performance impact (thoughts appreciated). ## Additional Info - [x] For performance it would be great to remove the cloning of full blocks when converting them to blinded blocks to write to disk. I'm going to experiment with a `put_block` API that takes the block by value, breaks it into a blinded block and a payload, stores the blinded block, and then re-assembles the full block for the caller. - [x] We should measure the latency of blocks-by-root and blocks-by-range responses. - [x] We should add integration tests that stress the payload reconstruction (basic tests done, issue for more extensive tests: https://github.com/sigp/lighthouse/issues/3159) - [x] We should (manually) test the schema v9 migration from several prior versions, particularly as blocks have changed on disk and some migrations rely on being able to load blocks. Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-05-12 00:42:17 +00:00
pub async fn get_block(
Update to frozen spec ❄️ (v0.8.1) (#444) * types: first updates for v0.8 * state_processing: epoch processing v0.8.0 * state_processing: block processing v0.8.0 * tree_hash_derive: support generics in SignedRoot * types v0.8: update to use ssz_types * state_processing v0.8: use ssz_types * ssz_types: add bitwise methods and from_elem * types: fix v0.8 FIXMEs * ssz_types: add bitfield shift_up * ssz_types: iterators and DerefMut for VariableList * types,state_processing: use VariableList * ssz_types: fix BitVector Decode impl Fixed a typo in the implementation of ssz::Decode for BitVector, which caused it to be considered variable length! * types: fix test modules for v0.8 update * types: remove slow type-level arithmetic * state_processing: fix tests for v0.8 * op_pool: update for v0.8 * ssz_types: Bitfield difference length-independent Allow computing the difference of two bitfields of different lengths. * Implement compact committee support * epoch_processing: committee & active index roots * state_processing: genesis state builder v0.8 * state_processing: implement v0.8.1 * Further improve tree_hash * Strip examples, tests from cached_tree_hash * Update TreeHash, un-impl CachedTreeHash * Update bitfield TreeHash, un-impl CachedTreeHash * Update FixedLenVec TreeHash, unimpl CachedTreeHash * Update update tree_hash_derive for new TreeHash * Fix TreeHash, un-impl CachedTreeHash for ssz_types * Remove fixed_len_vec, ssz benches SSZ benches relied upon fixed_len_vec -- it is easier to just delete them and rebuild them later (when necessary) * Remove boolean_bitfield crate * Fix fake_crypto BLS compile errors * Update ef_tests for new v.8 type params * Update ef_tests submodule to v0.8.1 tag * Make fixes to support parsing ssz ef_tests * `compact_committee...` to `compact_committees...` * Derive more traits for `CompactCommittee` * Flip bitfield byte-endianness * Fix tree_hash for bitfields * Modify CLI output for ef_tests * Bump ssz crate version * Update ssz_types doc comment * Del cached tree hash tests from ssz_static tests * Tidy SSZ dependencies * Rename ssz_types crate to eth2_ssz_types * validator_client: update for v0.8 * ssz_types: update union/difference for bit order swap * beacon_node: update for v0.8, EthSpec * types: disable cached tree hash, update min spec * state_processing: fix slot bug in committee update * tests: temporarily disable fork choice harness test See #447 * committee cache: prevent out-of-bounds access In the case where we tried to access the committee of a shard that didn't have a committee in the current epoch, we were accessing elements beyond the end of the shuffling vector and panicking! This commit adds a check to make the failure safe and explicit. * fix bug in get_indexed_attestation and simplify There was a bug in our implementation of get_indexed_attestation whereby incorrect "committee indices" were used to index into the custody bitfield. The bug was only observable in the case where some bits of the custody bitfield were set to 1. The implementation has been simplified to remove the bug, and a test added. * state_proc: workaround for compact committees bug https://github.com/ethereum/eth2.0-specs/issues/1315 * v0.8: updates to make the EF tests pass * Remove redundant max operation checks. * Always supply both messages when checking attestation signatures -- allowing verification of an attestation with no signatures. * Swap the order of the fork and domain constant in `get_domain`, to match the spec. * rustfmt * ef_tests: add new epoch processing tests * Integrate v0.8 into master (compiles) * Remove unused crates, fix clippy lints * Replace v0.6.3 tags w/ v0.8.1 * Remove old comment * Ensure lmd ghost tests only run in release * Update readme
2019-07-30 02:44:51 +00:00
&self,
block_root: &Hash256,
Update to Spec v0.10 (#817) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Disable/delete two outdated tests * Bump eth1 default vote warn to error * Delete outdated eth1 test Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-02-10 23:19:36 +00:00
) -> Result<Option<SignedBeaconBlock<T::EthSpec>>, Error> {
Separate execution payloads in the DB (#3157) ## Proposed Changes Reduce post-merge disk usage by not storing finalized execution payloads in Lighthouse's database. :warning: **This is achieved in a backwards-incompatible way for networks that have already merged** :warning:. Kiln users and shadow fork enjoyers will be unable to downgrade after running the code from this PR. The upgrade migration may take several minutes to run, and can't be aborted after it begins. The main changes are: - New column in the database called `ExecPayload`, keyed by beacon block root. - The `BeaconBlock` column now stores blinded blocks only. - Lots of places that previously used full blocks now use blinded blocks, e.g. analytics APIs, block replay in the DB, etc. - On finalization: - `prune_abanonded_forks` deletes non-canonical payloads whilst deleting non-canonical blocks. - `migrate_db` deletes finalized canonical payloads whilst deleting finalized states. - Conversions between blinded and full blocks are implemented in a compositional way, duplicating some work from Sean's PR #3134. - The execution layer has a new `get_payload_by_block_hash` method that reconstructs a payload using the EE's `eth_getBlockByHash` call. - I've tested manually that it works on Kiln, using Geth and Nethermind. - This isn't necessarily the most efficient method, and new engine APIs are being discussed to improve this: https://github.com/ethereum/execution-apis/pull/146. - We're depending on the `ethers` master branch, due to lots of recent changes. We're also using a workaround for https://github.com/gakonst/ethers-rs/issues/1134. - Payload reconstruction is used in the HTTP API via `BeaconChain::get_block`, which is now `async`. Due to the `async` fn, the `blocking_json` wrapper has been removed. - Payload reconstruction is used in network RPC to serve blocks-by-{root,range} responses. Here the `async` adjustment is messier, although I think I've managed to come up with a reasonable compromise: the handlers take the `SendOnDrop` by value so that they can drop it on _task completion_ (after the `fn` returns). Still, this is introducing disk reads onto core executor threads, which may have a negative performance impact (thoughts appreciated). ## Additional Info - [x] For performance it would be great to remove the cloning of full blocks when converting them to blinded blocks to write to disk. I'm going to experiment with a `put_block` API that takes the block by value, breaks it into a blinded block and a payload, stores the blinded block, and then re-assembles the full block for the caller. - [x] We should measure the latency of blocks-by-root and blocks-by-range responses. - [x] We should add integration tests that stress the payload reconstruction (basic tests done, issue for more extensive tests: https://github.com/sigp/lighthouse/issues/3159) - [x] We should (manually) test the schema v9 migration from several prior versions, particularly as blocks have changed on disk and some migrations rely on being able to load blocks. Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-05-12 00:42:17 +00:00
// Load block from database, returning immediately if we have the full block w payload
// stored.
let blinded_block = match self.store.try_get_full_block(block_root)? {
Some(DatabaseBlock::Full(block)) => return Ok(Some(block)),
Some(DatabaseBlock::Blinded(block)) => block,
None => return Ok(None),
};
// If we only have a blinded block, load the execution payload from the EL.
let block_message = blinded_block.message();
let execution_payload_header = &block_message
.execution_payload()
.map_err(|_| Error::BlockVariantLacksExecutionPayload(*block_root))?
.execution_payload_header;
let exec_block_hash = execution_payload_header.block_hash;
let execution_payload = self
.execution_layer
.as_ref()
.ok_or(Error::ExecutionLayerMissing)?
.get_payload_by_block_hash(exec_block_hash)
.await
.map_err(|e| Error::ExecutionLayerErrorPayloadReconstruction(exec_block_hash, e))?
.ok_or(Error::BlockHashMissingFromExecutionLayer(exec_block_hash))?;
// Verify payload integrity.
let header_from_payload = ExecutionPayloadHeader::from(&execution_payload);
if header_from_payload != *execution_payload_header {
for txn in &execution_payload.transactions {
debug!(
self.log,
"Reconstructed txn";
"bytes" => format!("0x{}", hex::encode(&**txn)),
);
}
return Err(Error::InconsistentPayloadReconstructed {
slot: blinded_block.slot(),
exec_block_hash,
canonical_payload_root: execution_payload_header.tree_hash_root(),
reconstructed_payload_root: header_from_payload.tree_hash_root(),
canonical_transactions_root: execution_payload_header.transactions_root,
reconstructed_transactions_root: header_from_payload.transactions_root,
});
}
// Add the payload to the block to form a full block.
blinded_block
.try_into_full_block(Some(execution_payload))
.ok_or(Error::AddPayloadLogicError)
.map(Some)
}
pub fn get_blinded_block(
&self,
block_root: &Hash256,
) -> Result<Option<SignedBlindedBeaconBlock<T::EthSpec>>, Error> {
Ok(self.store.get_blinded_block(block_root)?)
2019-03-23 02:23:44 +00:00
}
/// Returns the state at the given root, if any.
///
/// ## Errors
///
/// May return a database error.
pub fn get_state(
&self,
state_root: &Hash256,
slot: Option<Slot>,
) -> Result<Option<BeaconState<T::EthSpec>>, Error> {
Ok(self.store.get_state(state_root, slot)?)
}
/// Return the sync committee at `slot + 1` from the canonical chain.
///
/// This is useful when dealing with sync committee messages, because messages are signed
/// and broadcast one slot prior to the slot of the sync committee (which is relevant at
/// sync committee period boundaries).
pub fn sync_committee_at_next_slot(
&self,
slot: Slot,
) -> Result<Arc<SyncCommittee<T::EthSpec>>, Error> {
let epoch = slot.safe_add(1)?.epoch(T::EthSpec::slots_per_epoch());
self.sync_committee_at_epoch(epoch)
}
/// Return the sync committee at `epoch` from the canonical chain.
pub fn sync_committee_at_epoch(
&self,
epoch: Epoch,
) -> Result<Arc<SyncCommittee<T::EthSpec>>, Error> {
// Try to read a committee from the head. This will work most of the time, but will fail
// for faraway committees, or if there are skipped slots at the transition to Altair.
let spec = &self.spec;
let committee_from_head =
self.with_head(
|head| match head.beacon_state.get_built_sync_committee(epoch, spec) {
Ok(committee) => Ok(Some(committee.clone())),
Err(BeaconStateError::SyncCommitteeNotKnown { .. })
| Err(BeaconStateError::IncorrectStateVariant) => Ok(None),
Err(e) => Err(Error::from(e)),
},
)?;
if let Some(committee) = committee_from_head {
Ok(committee)
} else {
// Slow path: load a state (or advance the head).
let sync_committee_period = epoch.sync_committee_period(spec)?;
let committee = self
.state_for_sync_committee_period(sync_committee_period)?
.get_built_sync_committee(epoch, spec)?
.clone();
Ok(committee)
}
}
/// Load a state suitable for determining the sync committee for the given period.
///
/// Specifically, the state at the start of the *previous* sync committee period.
///
/// This is sufficient for historical duties, and efficient in the case where the head
/// is lagging the current period and we need duties for the next period (because we only
/// have to transition the head to start of the current period).
///
/// We also need to ensure that the load slot is after the Altair fork.
///
/// **WARNING**: the state returned will have dummy state roots. It should only be used
/// for its sync committees (determining duties, etc).
pub fn state_for_sync_committee_period(
&self,
sync_committee_period: u64,
) -> Result<BeaconState<T::EthSpec>, Error> {
let altair_fork_epoch = self
.spec
.altair_fork_epoch
.ok_or(Error::AltairForkDisabled)?;
let load_slot = std::cmp::max(
self.spec.epochs_per_sync_committee_period * sync_committee_period.saturating_sub(1),
altair_fork_epoch,
)
.start_slot(T::EthSpec::slots_per_epoch());
self.state_at_slot(load_slot, StateSkipConfig::WithoutStateRoots)
}
Prepare for public testnet (#628) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Update to use v0.9.2 version of deposit contract * Add efforts to automate eth1 testnet deployment * Fix lcli testnet deployer * Modify bn CLI to parse eth2_testnet_dir * Progress with account_manager deposit tools * Make account manager submit deposits * Add password option for submitting deposits * Allow custom deposit amount * Add long names to lcli clap * Add password option to lcli deploy command * Fix minor bugs whilst testing * Address Michael's comments * Add refund-deposit-contract to lcli * Use time instead of skip count for denying long skips * Improve logging for eth1 * Fix bug with validator services exiting on error * Drop the block cache after genesis * Modify eth1 testnet config * Improve eth1 logging * Make validator wait until genesis time * Fix bug in eth1 voting * Add more logging to eth1 voting * Handle errors in eth1 http module * Set SECONDS_PER_DAY to sensible minimum * Shorten delay before testnet start * Ensure eth1 block is produced without any votes * Improve eth1 logging * Fix broken tests in eth1 * Tidy code in rest_api * Fix failing test in deposit_contract * Make CLI args more consistent * Change validator/duties endpoint * Add time-based skip slot limiting * Add new error type missed in previous commit * Add log when waiting for genesis * Refactor beacon node CLI * Remove unused dep * Add lcli eth1-genesis command * Fix bug in master merge * Apply clippy lints to beacon node * Add support for YamlConfig in Eth2TestnetDir * Upgrade tesnet deposit contract version * Remove unnecessary logging and correct formatting * Add a hardcoded eth2 testnet config * Ensure http server flag works. Overwrite configs with flags. * Ensure boot nodes are loaded from testnet dir * Fix account manager CLI bugs * Fix bugs with beacon node cli * Allow testnet dir without boot nodes * Write genesis state as SSZ * Remove ---/n from the start of testnet_dir files * Set default libp2p address * Tidy account manager CLI, add logging * Add check to see if testnet dir exists * Apply reviewers suggestions * Add HeadTracker struct * Add fork choice persistence * Shorten slot time for simulator * Add the /beacon/heads API endpoint * Update hardcoded testnet * Add tests for BeaconChain persistence + fix bugs * Extend BeaconChain persistence testing * Ensure chain is finalized b4 persistence tests * Ensure boot_enr.yaml is include in binary * Refactor beacon_chain_sim * Move files about in beacon sim * Update beacon_chain_sim * Fix bug with deposit inclusion * Increase log in genesis service, fix todo * Tidy sim, fix broken rest_api tests * Fix more broken tests * Update testnet * Fix broken rest api test * Tidy account manager CLI * Use tempdir for account manager * Stop hardcoded testnet dir from creating dir * Rename Eth2TestnetDir to Eth2TestnetConfig * Change hardcoded -> hard_coded * Tidy account manager * Add log to account manager * Tidy, ensure head tracker is loaded from disk * Tidy beacon chain builder * Tidy eth1_chain * Adds log support for simulator * Revert "Adds log support for simulator" This reverts commit ec77c66a052350f551db145cf20f213823428dd3. * Adds log support for simulator * Tidy after self-review * Change default log level * Address Michael's delicious PR comments * Fix off-by-one in tests
2019-12-03 04:28:57 +00:00
/// Returns the current heads of the `BeaconChain`. For the canonical head, see `Self::head`.
///
/// Returns `(block_root, block_slot)`.
pub fn heads(&self) -> Vec<(Hash256, Slot)> {
self.head_tracker.heads()
}
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
pub fn knows_head(&self, block_hash: &SignedBeaconBlockHash) -> bool {
self.head_tracker.contains_head((*block_hash).into())
}
/// Returns the `BeaconState` at the given slot.
///
Prepare for public testnet (#628) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Update to use v0.9.2 version of deposit contract * Add efforts to automate eth1 testnet deployment * Fix lcli testnet deployer * Modify bn CLI to parse eth2_testnet_dir * Progress with account_manager deposit tools * Make account manager submit deposits * Add password option for submitting deposits * Allow custom deposit amount * Add long names to lcli clap * Add password option to lcli deploy command * Fix minor bugs whilst testing * Address Michael's comments * Add refund-deposit-contract to lcli * Use time instead of skip count for denying long skips * Improve logging for eth1 * Fix bug with validator services exiting on error * Drop the block cache after genesis * Modify eth1 testnet config * Improve eth1 logging * Make validator wait until genesis time * Fix bug in eth1 voting * Add more logging to eth1 voting * Handle errors in eth1 http module * Set SECONDS_PER_DAY to sensible minimum * Shorten delay before testnet start * Ensure eth1 block is produced without any votes * Improve eth1 logging * Fix broken tests in eth1 * Tidy code in rest_api * Fix failing test in deposit_contract * Make CLI args more consistent * Change validator/duties endpoint * Add time-based skip slot limiting * Add new error type missed in previous commit * Add log when waiting for genesis * Refactor beacon node CLI * Remove unused dep * Add lcli eth1-genesis command * Fix bug in master merge * Apply clippy lints to beacon node * Add support for YamlConfig in Eth2TestnetDir * Upgrade tesnet deposit contract version * Remove unnecessary logging and correct formatting * Add a hardcoded eth2 testnet config * Ensure http server flag works. Overwrite configs with flags. * Ensure boot nodes are loaded from testnet dir * Fix account manager CLI bugs * Fix bugs with beacon node cli * Allow testnet dir without boot nodes * Write genesis state as SSZ * Remove ---/n from the start of testnet_dir files * Set default libp2p address * Tidy account manager CLI, add logging * Add check to see if testnet dir exists * Apply reviewers suggestions * Add HeadTracker struct * Add fork choice persistence * Shorten slot time for simulator * Add the /beacon/heads API endpoint * Update hardcoded testnet * Add tests for BeaconChain persistence + fix bugs * Extend BeaconChain persistence testing * Ensure chain is finalized b4 persistence tests * Ensure boot_enr.yaml is include in binary * Refactor beacon_chain_sim * Move files about in beacon sim * Update beacon_chain_sim * Fix bug with deposit inclusion * Increase log in genesis service, fix todo * Tidy sim, fix broken rest_api tests * Fix more broken tests * Update testnet * Fix broken rest api test * Tidy account manager CLI * Use tempdir for account manager * Stop hardcoded testnet dir from creating dir * Rename Eth2TestnetDir to Eth2TestnetConfig * Change hardcoded -> hard_coded * Tidy account manager * Add log to account manager * Tidy, ensure head tracker is loaded from disk * Tidy beacon chain builder * Tidy eth1_chain * Adds log support for simulator * Revert "Adds log support for simulator" This reverts commit ec77c66a052350f551db145cf20f213823428dd3. * Adds log support for simulator * Tidy after self-review * Change default log level * Address Michael's delicious PR comments * Fix off-by-one in tests
2019-12-03 04:28:57 +00:00
/// Returns `None` when the state is not found in the database or there is an error skipping
/// to a future state.
Faster attestation production (#838) * Start adding interop genesis state to lcli * Use more efficient method to generate genesis state * Remove duplicate int_to_bytes32 * Add lcli command to change state genesis time * Add option to allow VC to start with unsynced BN * Set VC to do parallel key loading * Don't default to dummy eth1 backend * Add endpoint to dump operation pool * Add metrics for op pool * Remove state clone for slot notifier * Add mem size approximation for tree hash cache * Avoid cloning tree hash when getting head * Avoid cloning tree hash when getting head * Add working arena-based cached tree hash * Add another benchmark * Add pre-allocation for caches * Make cache nullable * Fix bugs in cache tree hash * Add validator tree hash optimization * Optimize hash_concat * Make hash32_concat return fixed-len array * Fix failing API tests * Add new beacon state cache struct * Add validator-specific cache * Separate list and values arenas * Add parallel validator registry hashing * Remove MultiTreeHashCache * Remove cached tree hash macro * Fix failing tree hash test * Address Michael's comments * Add CachedTreeHash impl for ef tests * Fix messy merge conflict * Optimize attestation production * Add first basic optimizations * Fix SlotOutOfBounds error * Resolved missed merge conflicts * Fix another missed merge conflict * Fix more merge conflict issues * Add `StateSkipConfig` * Fix test compile errors * Add failing test * Fix bug, make tests pass * Add comment * Delete unused function * Replace deleted comment
2020-03-04 06:10:22 +00:00
pub fn state_at_slot(
&self,
slot: Slot,
config: StateSkipConfig,
) -> Result<BeaconState<T::EthSpec>, Error> {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let head_state = self.head_beacon_state_cloned();
match slot.cmp(&head_state.slot()) {
Ordering::Equal => Ok(head_state),
Ordering::Greater => {
if slot > head_state.slot() + T::EthSpec::slots_per_epoch() {
warn!(
self.log,
"Skipping more than an epoch";
"head_slot" => head_state.slot(),
"request_slot" => slot
)
}
let start_slot = head_state.slot();
let task_start = Instant::now();
let max_task_runtime = Duration::from_secs(self.spec.seconds_per_slot);
let head_state_slot = head_state.slot();
let mut state = head_state;
Faster attestation production (#838) * Start adding interop genesis state to lcli * Use more efficient method to generate genesis state * Remove duplicate int_to_bytes32 * Add lcli command to change state genesis time * Add option to allow VC to start with unsynced BN * Set VC to do parallel key loading * Don't default to dummy eth1 backend * Add endpoint to dump operation pool * Add metrics for op pool * Remove state clone for slot notifier * Add mem size approximation for tree hash cache * Avoid cloning tree hash when getting head * Avoid cloning tree hash when getting head * Add working arena-based cached tree hash * Add another benchmark * Add pre-allocation for caches * Make cache nullable * Fix bugs in cache tree hash * Add validator tree hash optimization * Optimize hash_concat * Make hash32_concat return fixed-len array * Fix failing API tests * Add new beacon state cache struct * Add validator-specific cache * Separate list and values arenas * Add parallel validator registry hashing * Remove MultiTreeHashCache * Remove cached tree hash macro * Fix failing tree hash test * Address Michael's comments * Add CachedTreeHash impl for ef tests * Fix messy merge conflict * Optimize attestation production * Add first basic optimizations * Fix SlotOutOfBounds error * Resolved missed merge conflicts * Fix another missed merge conflict * Fix more merge conflict issues * Add `StateSkipConfig` * Fix test compile errors * Add failing test * Fix bug, make tests pass * Add comment * Delete unused function * Replace deleted comment
2020-03-04 06:10:22 +00:00
let skip_state_root = match config {
StateSkipConfig::WithStateRoots => None,
StateSkipConfig::WithoutStateRoots => Some(Hash256::zero()),
};
while state.slot() < slot {
// Do not allow and forward state skip that takes longer than the maximum task duration.
//
// This is a protection against nodes doing too much work when they're not synced
// to a chain.
if task_start + max_task_runtime < Instant::now() {
return Err(Error::StateSkipTooLarge {
start_slot,
requested_slot: slot,
max_task_runtime,
});
}
// Note: supplying some `state_root` when it is known would be a cheap and easy
// optimization.
Faster attestation production (#838) * Start adding interop genesis state to lcli * Use more efficient method to generate genesis state * Remove duplicate int_to_bytes32 * Add lcli command to change state genesis time * Add option to allow VC to start with unsynced BN * Set VC to do parallel key loading * Don't default to dummy eth1 backend * Add endpoint to dump operation pool * Add metrics for op pool * Remove state clone for slot notifier * Add mem size approximation for tree hash cache * Avoid cloning tree hash when getting head * Avoid cloning tree hash when getting head * Add working arena-based cached tree hash * Add another benchmark * Add pre-allocation for caches * Make cache nullable * Fix bugs in cache tree hash * Add validator tree hash optimization * Optimize hash_concat * Make hash32_concat return fixed-len array * Fix failing API tests * Add new beacon state cache struct * Add validator-specific cache * Separate list and values arenas * Add parallel validator registry hashing * Remove MultiTreeHashCache * Remove cached tree hash macro * Fix failing tree hash test * Address Michael's comments * Add CachedTreeHash impl for ef tests * Fix messy merge conflict * Optimize attestation production * Add first basic optimizations * Fix SlotOutOfBounds error * Resolved missed merge conflicts * Fix another missed merge conflict * Fix more merge conflict issues * Add `StateSkipConfig` * Fix test compile errors * Add failing test * Fix bug, make tests pass * Add comment * Delete unused function * Replace deleted comment
2020-03-04 06:10:22 +00:00
match per_slot_processing(&mut state, skip_state_root, &self.spec) {
Stable futures (#879) * Port eth1 lib to use stable futures * Port eth1_test_rig to stable futures * Port eth1 tests to stable futures * Port genesis service to stable futures * Port genesis tests to stable futures * Port beacon_chain to stable futures * Port lcli to stable futures * Fix eth1_test_rig (#1014) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Update hashmap hashset to stable futures * Adds panic test to hashset delay * Port remote_beacon_node to stable futures * Fix lcli merge conflicts * Non rpc stuff compiles * protocol.rs compiles * Port websockets, timer and notifier to stable futures (#1035) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Port remote_beacon_node to stable futures * Partial eth2-libp2p stable future upgrade * Finished first round of fighting RPC types * Further progress towards porting eth2-libp2p adds caching to discovery * Update behaviour * RPC handler to stable futures * Update RPC to master libp2p * Network service additions * Fix the fallback transport construction (#1102) * Correct warning * Remove hashmap delay * Compiling version of eth2-libp2p * Update all crates versions * Fix conversion function and add tests (#1113) * Port validator_client to stable futures (#1114) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Revert "Update hashmap hashset to stable futures" This reverts commit d432378a3cc5cd67fc29c0b15b96b886c1323554. * Revert "Adds panic test to hashset delay" This reverts commit 281502396fc5b90d9c421a309c2c056982c9525b. * Ported attestation_service * Ported duties_service * Ported fork_service * More ports * Port block_service * Minor fixes * VC compiles * Update TODOS * Borrow self where possible * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test * Re-Fix attestation and duties service * Minor fixes Co-authored-by: Paul Hauner <paul@paulhauner.com> * Network crate update to stable futures * Port account_manager to stable futures (#1121) * Port account_manager to stable futures * Run async fns in tokio environment * Port rest_api crate to stable futures (#1118) * Port rest_api lib to stable futures * Reduce tokio features * Update notifier to stable futures * Builder update * Further updates * Convert self referential async functions * stable futures fixes (#1124) * Fix eth1 update functions * Fix genesis and client * Fix beacon node lib * Return appropriate runtimes from environment * Fix test rig * Refactor eth1 service update * Upgrade simulator to stable futures * Lighthouse compiles on stable futures * Remove println debugging statement * Update libp2p service, start rpc test upgrade * Update network crate for new libp2p * Update tokio::codec to futures_codec (#1128) * Further work towards RPC corrections * Correct http timeout and network service select * Use tokio runtime for libp2p * Revert "Update tokio::codec to futures_codec (#1128)" This reverts commit e57aea924acf5cbabdcea18895ac07e38a425ed7. * Upgrade RPC libp2p tests * Upgrade secio fallback test * Upgrade gossipsub examples * Clean up RPC protocol * Test fixes (#1133) * Correct websocket timeout and run on os thread * Fix network test * Clean up PR * Correct tokio tcp move attestation service tests * Upgrade attestation service tests * Correct network test * Correct genesis test * Test corrections * Log info when block is received * Modify logs and update attester service events * Stable futures: fixes to vc, eth1 and account manager (#1142) * Add local testnet scripts * Remove whiteblock script * Rename local testnet script * Move spawns onto handle * Fix VC panic * Initial fix to block production issue * Tidy block producer fix * Tidy further * Add local testnet clean script * Run cargo fmt * Tidy duties service * Tidy fork service * Tidy ForkService * Tidy AttestationService * Tidy notifier * Ensure await is not suppressed in eth1 * Ensure await is not suppressed in account_manager * Use .ok() instead of .unwrap_or(()) * RPC decoding test for proto * Update discv5 and eth2-libp2p deps * Fix lcli double runtime issue (#1144) * Handle stream termination and dialing peer errors * Correct peer_info variant types * Remove unnecessary warnings * Handle subnet unsubscription removal and improve logigng * Add logs around ping * Upgrade discv5 and improve logging * Handle peer connection status for multiple connections * Improve network service logging * Improve logging around peer manager * Upgrade swarm poll centralise peer management * Identify clients on error * Fix `remove_peer` in sync (#1150) * remove_peer removes from all chains * Remove logs * Fix early return from loop * Improved logging, fix panic * Partially correct tests * Stable futures: Vc sync (#1149) * Improve syncing heuristic * Add comments * Use safer method for tolerance * Fix tests * Stable futures: Fix VC bug, update agg pool, add more metrics (#1151) * Expose epoch processing summary * Expose participation metrics to prometheus * Switch to f64 * Reduce precision * Change precision * Expose observed attesters metrics * Add metrics for agg/unagg attn counts * Add metrics for gossip rx * Add metrics for gossip tx * Adds ignored attns to prom * Add attestation timing * Add timer for aggregation pool sig agg * Add write lock timer for agg pool * Add more metrics to agg pool * Change map lock code * Add extra metric to agg pool * Change lock handling in agg pool * Change .write() to .read() * Add another agg pool timer * Fix for is_aggregator * Fix pruning bug Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-05-17 11:16:48 +00:00
Ok(_) => (),
Err(e) => {
warn!(
self.log,
"Unable to load state at slot";
"error" => ?e,
"head_slot" => head_state_slot,
"requested_slot" => slot
);
return Err(Error::NoStateForSlot(slot));
}
};
}
Ok(state)
}
Ordering::Less => {
let state_root =
process_results(self.forwards_iter_state_roots_until(slot, slot)?, |iter| {
iter.take_while(|(_, current_slot)| *current_slot >= slot)
.find(|(_, current_slot)| *current_slot == slot)
.map(|(root, _slot)| root)
})?
.ok_or(Error::NoStateForSlot(slot))?;
Ok(self
.get_state(&state_root, Some(slot))?
.ok_or(Error::NoStateForSlot(slot))?)
2019-06-20 00:59:19 +00:00
}
}
}
/// Returns the `BeaconState` the current slot (viz., `self.slot()`).
///
/// - A reference to the head state (note: this keeps a read lock on the head, try to use
/// sparingly).
/// - The head state, but with skipped slots (for states later than the head).
///
/// Returns `None` when there is an error skipping to a future state or the slot clock cannot
/// be read.
2019-09-11 02:42:07 +00:00
pub fn wall_clock_state(&self) -> Result<BeaconState<T::EthSpec>, Error> {
Faster attestation production (#838) * Start adding interop genesis state to lcli * Use more efficient method to generate genesis state * Remove duplicate int_to_bytes32 * Add lcli command to change state genesis time * Add option to allow VC to start with unsynced BN * Set VC to do parallel key loading * Don't default to dummy eth1 backend * Add endpoint to dump operation pool * Add metrics for op pool * Remove state clone for slot notifier * Add mem size approximation for tree hash cache * Avoid cloning tree hash when getting head * Avoid cloning tree hash when getting head * Add working arena-based cached tree hash * Add another benchmark * Add pre-allocation for caches * Make cache nullable * Fix bugs in cache tree hash * Add validator tree hash optimization * Optimize hash_concat * Make hash32_concat return fixed-len array * Fix failing API tests * Add new beacon state cache struct * Add validator-specific cache * Separate list and values arenas * Add parallel validator registry hashing * Remove MultiTreeHashCache * Remove cached tree hash macro * Fix failing tree hash test * Address Michael's comments * Add CachedTreeHash impl for ef tests * Fix messy merge conflict * Optimize attestation production * Add first basic optimizations * Fix SlotOutOfBounds error * Resolved missed merge conflicts * Fix another missed merge conflict * Fix more merge conflict issues * Add `StateSkipConfig` * Fix test compile errors * Add failing test * Fix bug, make tests pass * Add comment * Delete unused function * Replace deleted comment
2020-03-04 06:10:22 +00:00
self.state_at_slot(self.slot()?, StateSkipConfig::WithStateRoots)
}
/// Returns the validator index (if any) for the given public key.
///
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
/// ## Notes
///
/// This query uses the `validator_pubkey_cache` which contains _all_ validators ever seen,
/// even if those validators aren't included in the head state. It is important to remember
/// that just because a validator exists here, it doesn't necessarily exist in all
/// `BeaconStates`.
///
/// ## Errors
///
/// May return an error if acquiring a read-lock on the `validator_pubkey_cache` times out.
pub fn validator_index(&self, pubkey: &PublicKeyBytes) -> Result<Option<usize>, Error> {
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
let pubkey_cache = self
.validator_pubkey_cache
.try_read_for(VALIDATOR_PUBKEY_CACHE_LOCK_TIMEOUT)
.ok_or(Error::ValidatorPubkeyCacheLockTimeout)?;
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
Ok(pubkey_cache.get_index(pubkey))
}
/// Return the validator indices of all public keys fetched from an iterator.
///
/// If any public key doesn't belong to a known validator then an error will be returned.
/// We could consider relaxing this by returning `Vec<Option<usize>>` in future.
pub fn validator_indices<'a>(
&self,
validator_pubkeys: impl Iterator<Item = &'a PublicKeyBytes>,
) -> Result<Vec<u64>, Error> {
let pubkey_cache = self
.validator_pubkey_cache
.try_read_for(VALIDATOR_PUBKEY_CACHE_LOCK_TIMEOUT)
.ok_or(Error::ValidatorPubkeyCacheLockTimeout)?;
validator_pubkeys
.map(|pubkey| {
pubkey_cache
.get_index(pubkey)
.map(|id| id as u64)
.ok_or(Error::ValidatorPubkeyUnknown(*pubkey))
})
.collect()
}
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
/// Returns the validator pubkey (if any) for the given validator index.
///
/// ## Notes
///
/// This query uses the `validator_pubkey_cache` which contains _all_ validators ever seen,
/// even if those validators aren't included in the head state. It is important to remember
/// that just because a validator exists here, it doesn't necessarily exist in all
/// `BeaconStates`.
///
/// ## Errors
///
/// May return an error if acquiring a read-lock on the `validator_pubkey_cache` times out.
pub fn validator_pubkey(&self, validator_index: usize) -> Result<Option<PublicKey>, Error> {
let pubkey_cache = self
.validator_pubkey_cache
.try_read_for(VALIDATOR_PUBKEY_CACHE_LOCK_TIMEOUT)
.ok_or(Error::ValidatorPubkeyCacheLockTimeout)?;
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
Ok(pubkey_cache.get(validator_index).cloned())
}
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
/// As per `Self::validator_pubkey`, but returns `PublicKeyBytes`.
pub fn validator_pubkey_bytes(
&self,
validator_index: usize,
) -> Result<Option<PublicKeyBytes>, Error> {
let pubkey_cache = self
.validator_pubkey_cache
.try_read_for(VALIDATOR_PUBKEY_CACHE_LOCK_TIMEOUT)
.ok_or(Error::ValidatorPubkeyCacheLockTimeout)?;
Ok(pubkey_cache.get_pubkey_bytes(validator_index).copied())
}
/// As per `Self::validator_pubkey_bytes` but will resolve multiple indices at once to avoid
/// bouncing the read-lock on the pubkey cache.
///
/// Returns a map that may have a length less than `validator_indices.len()` if some indices
/// were unable to be resolved.
pub fn validator_pubkey_bytes_many(
&self,
validator_indices: &[usize],
) -> Result<HashMap<usize, PublicKeyBytes>, Error> {
let pubkey_cache = self
.validator_pubkey_cache
.try_read_for(VALIDATOR_PUBKEY_CACHE_LOCK_TIMEOUT)
.ok_or(Error::ValidatorPubkeyCacheLockTimeout)?;
let mut map = HashMap::with_capacity(validator_indices.len());
for &validator_index in validator_indices {
if let Some(pubkey) = pubkey_cache.get_pubkey_bytes(validator_index) {
map.insert(validator_index, *pubkey);
}
}
Ok(map)
}
/// Returns the block canonical root of the current canonical chain at a given slot, starting from the given state.
///
/// Returns `None` if the given slot doesn't exist in the chain.
pub fn root_at_slot_from_state(
&self,
target_slot: Slot,
beacon_block_root: Hash256,
state: &BeaconState<T::EthSpec>,
) -> Result<Option<Hash256>, Error> {
let iter = BlockRootsIterator::new(&self.store, state);
let iter_with_head = std::iter::once(Ok((beacon_block_root, state.slot())))
.chain(iter)
.map(|result| result.map_err(|e| e.into()));
process_results(iter_with_head, |mut iter| {
iter.find(|(_, slot)| *slot == target_slot)
.map(|(root, _)| root)
})
}
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
/// Returns the attestation duties for the given validator indices using the shuffling cache.
///
/// An error may be returned if `head_block_root` is a finalized block, this function is only
/// designed for operations at the head of the chain.
///
/// The returned `Vec` will have the same length as `validator_indices`, any
/// non-existing/inactive validators will have `None` values.
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
///
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
/// ## Notes
///
/// This function will try to use the shuffling cache to return the value. If the value is not
/// in the shuffling cache, it will be added. Care should be taken not to wash out the
/// shuffling cache with historical/useless values.
pub fn validator_attestation_duties(
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
&self,
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
validator_indices: &[u64],
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
epoch: Epoch,
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
head_block_root: Hash256,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
) -> Result<(Vec<Option<AttestationDuty>>, Hash256, ExecutionStatus), Error> {
Add `execution_optimistic` flag to HTTP responses (#3070) ## Issue Addressed #3031 ## Proposed Changes Updates the following API endpoints to conform with https://github.com/ethereum/beacon-APIs/pull/190 and https://github.com/ethereum/beacon-APIs/pull/196 - [x] `beacon/states/{state_id}/root` - [x] `beacon/states/{state_id}/fork` - [x] `beacon/states/{state_id}/finality_checkpoints` - [x] `beacon/states/{state_id}/validators` - [x] `beacon/states/{state_id}/validators/{validator_id}` - [x] `beacon/states/{state_id}/validator_balances` - [x] `beacon/states/{state_id}/committees` - [x] `beacon/states/{state_id}/sync_committees` - [x] `beacon/headers` - [x] `beacon/headers/{block_id}` - [x] `beacon/blocks/{block_id}` - [x] `beacon/blocks/{block_id}/root` - [x] `beacon/blocks/{block_id}/attestations` - [x] `debug/beacon/states/{state_id}` - [x] `debug/beacon/heads` - [x] `validator/duties/attester/{epoch}` - [x] `validator/duties/proposer/{epoch}` - [x] `validator/duties/sync/{epoch}` Updates the following Server-Sent Events: - [x] `events?topics=head` - [x] `events?topics=block` - [x] `events?topics=finalized_checkpoint` - [x] `events?topics=chain_reorg` ## Backwards Incompatible There is a very minor breaking change with the way the API now handles requests to `beacon/blocks/{block_id}/root` and `beacon/states/{state_id}/root` when `block_id` or `state_id` is the `Root` variant of `BlockId` and `StateId` respectively. Previously a request to a non-existent root would simply echo the root back to the requester: ``` curl "http://localhost:5052/eth/v1/beacon/states/0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa/root" {"data":{"root":"0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"}} ``` Now it will return a `404`: ``` curl "http://localhost:5052/eth/v1/beacon/blocks/0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa/root" {"code":404,"message":"NOT_FOUND: beacon block with root 0xaaaa…aaaa","stacktraces":[]} ``` In addition to this is the block root `0x0000000000000000000000000000000000000000000000000000000000000000` previously would return the genesis block. It will now return a `404`: ``` curl "http://localhost:5052/eth/v1/beacon/blocks/0x0000000000000000000000000000000000000000000000000000000000000000" {"code":404,"message":"NOT_FOUND: beacon block with root 0x0000…0000","stacktraces":[]} ``` ## Additional Info - `execution_optimistic` is always set, and will return `false` pre-Bellatrix. I am also open to the idea of doing something like `#[serde(skip_serializing_if = "Option::is_none")]`. - The value of `execution_optimistic` is set to `false` where possible. Any computation that is reliant on the `head` will simply use the `ExecutionStatus` of the head (unless the head block is pre-Bellatrix). Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-07-25 08:23:00 +00:00
let execution_status = self
.canonical_head
.fork_choice_read_lock()
.get_block_execution_status(&head_block_root)
.ok_or(Error::AttestationHeadNotInForkChoice(head_block_root))?;
let (duties, dependent_root) = self.with_committee_cache(
head_block_root,
epoch,
|committee_cache, dependent_root| {
let duties = validator_indices
.iter()
.map(|validator_index| {
let validator_index = *validator_index as usize;
committee_cache.get_attestation_duties(validator_index)
})
.collect();
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
Add `execution_optimistic` flag to HTTP responses (#3070) ## Issue Addressed #3031 ## Proposed Changes Updates the following API endpoints to conform with https://github.com/ethereum/beacon-APIs/pull/190 and https://github.com/ethereum/beacon-APIs/pull/196 - [x] `beacon/states/{state_id}/root` - [x] `beacon/states/{state_id}/fork` - [x] `beacon/states/{state_id}/finality_checkpoints` - [x] `beacon/states/{state_id}/validators` - [x] `beacon/states/{state_id}/validators/{validator_id}` - [x] `beacon/states/{state_id}/validator_balances` - [x] `beacon/states/{state_id}/committees` - [x] `beacon/states/{state_id}/sync_committees` - [x] `beacon/headers` - [x] `beacon/headers/{block_id}` - [x] `beacon/blocks/{block_id}` - [x] `beacon/blocks/{block_id}/root` - [x] `beacon/blocks/{block_id}/attestations` - [x] `debug/beacon/states/{state_id}` - [x] `debug/beacon/heads` - [x] `validator/duties/attester/{epoch}` - [x] `validator/duties/proposer/{epoch}` - [x] `validator/duties/sync/{epoch}` Updates the following Server-Sent Events: - [x] `events?topics=head` - [x] `events?topics=block` - [x] `events?topics=finalized_checkpoint` - [x] `events?topics=chain_reorg` ## Backwards Incompatible There is a very minor breaking change with the way the API now handles requests to `beacon/blocks/{block_id}/root` and `beacon/states/{state_id}/root` when `block_id` or `state_id` is the `Root` variant of `BlockId` and `StateId` respectively. Previously a request to a non-existent root would simply echo the root back to the requester: ``` curl "http://localhost:5052/eth/v1/beacon/states/0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa/root" {"data":{"root":"0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"}} ``` Now it will return a `404`: ``` curl "http://localhost:5052/eth/v1/beacon/blocks/0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa/root" {"code":404,"message":"NOT_FOUND: beacon block with root 0xaaaa…aaaa","stacktraces":[]} ``` In addition to this is the block root `0x0000000000000000000000000000000000000000000000000000000000000000` previously would return the genesis block. It will now return a `404`: ``` curl "http://localhost:5052/eth/v1/beacon/blocks/0x0000000000000000000000000000000000000000000000000000000000000000" {"code":404,"message":"NOT_FOUND: beacon block with root 0x0000…0000","stacktraces":[]} ``` ## Additional Info - `execution_optimistic` is always set, and will return `false` pre-Bellatrix. I am also open to the idea of doing something like `#[serde(skip_serializing_if = "Option::is_none")]`. - The value of `execution_optimistic` is set to `false` where possible. Any computation that is reliant on the `head` will simply use the `ExecutionStatus` of the head (unless the head block is pre-Bellatrix). Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-07-25 08:23:00 +00:00
Ok((duties, dependent_root))
},
)?;
Ok((duties, dependent_root, execution_status))
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
}
/// Returns an aggregated `Attestation`, if any, that has a matching `attestation.data`.
///
/// The attestation will be obtained from `self.naive_aggregation_pool`.
pub fn get_aggregated_attestation(
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
&self,
data: &AttestationData,
) -> Result<Option<Attestation<T::EthSpec>>, Error> {
if let Some(attestation) = self.naive_aggregation_pool.read().get(data) {
self.filter_optimistic_attestation(attestation)
.map(Option::Some)
} else {
Ok(None)
}
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
}
/// Returns an aggregated `Attestation`, if any, that has a matching
/// `attestation.data.tree_hash_root()`.
///
/// The attestation will be obtained from `self.naive_aggregation_pool`.
pub fn get_aggregated_attestation_by_slot_and_root(
&self,
slot: Slot,
attestation_data_root: &Hash256,
) -> Result<Option<Attestation<T::EthSpec>>, Error> {
if let Some(attestation) = self
.naive_aggregation_pool
.read()
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
.get_by_slot_and_root(slot, attestation_data_root)
{
self.filter_optimistic_attestation(attestation)
.map(Option::Some)
} else {
Ok(None)
}
}
/// Returns `Ok(attestation)` if the supplied `attestation` references a valid
/// `beacon_block_root`.
fn filter_optimistic_attestation(
&self,
attestation: Attestation<T::EthSpec>,
) -> Result<Attestation<T::EthSpec>, Error> {
let beacon_block_root = attestation.data.beacon_block_root;
match self
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
.canonical_head
.fork_choice_read_lock()
.get_block_execution_status(&beacon_block_root)
{
// The attestation references a block that is not in fork choice, it must be
// pre-finalization.
None => Err(Error::CannotAttestToFinalizedBlock { beacon_block_root }),
// The attestation references a fully valid `beacon_block_root`.
Some(execution_status) if execution_status.is_valid_or_irrelevant() => Ok(attestation),
// The attestation references a block that has not been verified by an EL (i.e. it
// is optimistic or invalid). Don't return the block, return an error instead.
Some(execution_status) => Err(Error::HeadBlockNotFullyVerified {
beacon_block_root,
execution_status,
}),
}
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
}
/// Return an aggregated `SyncCommitteeContribution` matching the given `root`.
pub fn get_aggregated_sync_committee_contribution(
&self,
sync_contribution_data: &SyncContributionData,
) -> Result<Option<SyncCommitteeContribution<T::EthSpec>>, Error> {
if let Some(contribution) = self
.naive_sync_aggregation_pool
.read()
.get(sync_contribution_data)
{
self.filter_optimistic_sync_committee_contribution(contribution)
.map(Option::Some)
} else {
Ok(None)
}
}
fn filter_optimistic_sync_committee_contribution(
&self,
contribution: SyncCommitteeContribution<T::EthSpec>,
) -> Result<SyncCommitteeContribution<T::EthSpec>, Error> {
let beacon_block_root = contribution.beacon_block_root;
match self
.canonical_head
.fork_choice_read_lock()
.get_block_execution_status(&beacon_block_root)
{
// The contribution references a block that is not in fork choice, it must be
// pre-finalization.
None => Err(Error::SyncContributionDataReferencesFinalizedBlock { beacon_block_root }),
// The contribution references a fully valid `beacon_block_root`.
Some(execution_status) if execution_status.is_valid_or_irrelevant() => Ok(contribution),
// The contribution references a block that has not been verified by an EL (i.e. it
// is optimistic or invalid). Don't return the block, return an error instead.
Some(execution_status) => Err(Error::HeadBlockNotFullyVerified {
beacon_block_root,
execution_status,
}),
}
}
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
/// Produce an unaggregated `Attestation` that is valid for the given `slot` and `index`.
///
/// The produced `Attestation` will not be valid until it has been signed by exactly one
/// validator that is in the committee for `slot` and `index` in the canonical chain.
Validator client refactor (#618) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Address Michael's comments
2019-11-25 04:48:24 +00:00
///
/// Always attests to the canonical chain.
///
/// ## Errors
///
/// May return an error if the `request_slot` is too far behind the head state.
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
pub fn produce_unaggregated_attestation(
Validator client refactor (#618) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Address Michael's comments
2019-11-25 04:48:24 +00:00
&self,
request_slot: Slot,
request_index: CommitteeIndex,
Validator client refactor (#618) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Address Michael's comments
2019-11-25 04:48:24 +00:00
) -> Result<Attestation<T::EthSpec>, Error> {
let _total_timer = metrics::start_timer(&metrics::ATTESTATION_PRODUCTION_SECONDS);
Validator client refactor (#618) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Address Michael's comments
2019-11-25 04:48:24 +00:00
Add early attester cache (#2872) ## Issue Addressed NA ## Proposed Changes Introduces a cache to attestation to produce atop blocks which will become the head, but are not fully imported (e.g., not inserted into the database). Whilst attesting to a block before it's imported is rather easy, if we're going to produce that attestation then we also need to be able to: 1. Verify that attestation. 1. Respond to RPC requests for the `beacon_block_root`. Attestation verification (1) is *partially* covered. Since we prime the shuffling cache before we insert the block into the early attester cache, we should be fine for all typical use-cases. However, it is possible that the cache is washed out before we've managed to insert the state into the database and then attestation verification will fail with a "missing beacon state"-type error. Providing the block via RPC (2) is also partially covered, since we'll check the database *and* the early attester cache when responding a blocks-by-root request. However, we'll still omit the block from blocks-by-range requests (until the block lands in the DB). I *think* this is fine, since there's no guarantee that we return all blocks for those responses. Another important consideration is whether or not the *parent* of the early attester block is available in the databse. If it were not, we might fail to respond to blocks-by-root request that are iterating backwards to collect a chain of blocks. I argue that *we will always have the parent of the early attester block in the database.* This is because we are holding the fork-choice write-lock when inserting the block into the early attester cache and we do not drop that until the block is in the database.
2022-01-11 01:35:55 +00:00
// The early attester cache will return `Some(attestation)` in the scenario where there is a
// block being imported that will become the head block, but that block has not yet been
// inserted into the database and set as `self.canonical_head`.
//
// In effect, the early attester cache prevents slow database IO from causing missed
// head/target votes.
//
// The early attester cache should never contain an optimistically imported block.
Add early attester cache (#2872) ## Issue Addressed NA ## Proposed Changes Introduces a cache to attestation to produce atop blocks which will become the head, but are not fully imported (e.g., not inserted into the database). Whilst attesting to a block before it's imported is rather easy, if we're going to produce that attestation then we also need to be able to: 1. Verify that attestation. 1. Respond to RPC requests for the `beacon_block_root`. Attestation verification (1) is *partially* covered. Since we prime the shuffling cache before we insert the block into the early attester cache, we should be fine for all typical use-cases. However, it is possible that the cache is washed out before we've managed to insert the state into the database and then attestation verification will fail with a "missing beacon state"-type error. Providing the block via RPC (2) is also partially covered, since we'll check the database *and* the early attester cache when responding a blocks-by-root request. However, we'll still omit the block from blocks-by-range requests (until the block lands in the DB). I *think* this is fine, since there's no guarantee that we return all blocks for those responses. Another important consideration is whether or not the *parent* of the early attester block is available in the databse. If it were not, we might fail to respond to blocks-by-root request that are iterating backwards to collect a chain of blocks. I argue that *we will always have the parent of the early attester block in the database.* This is because we are holding the fork-choice write-lock when inserting the block into the early attester cache and we do not drop that until the block is in the database.
2022-01-11 01:35:55 +00:00
match self
.early_attester_cache
.try_attest(request_slot, request_index, &self.spec)
{
// The cache matched this request, return the value.
Ok(Some(attestation)) => return Ok(attestation),
// The cache did not match this request, proceed with the rest of this function.
Ok(None) => (),
// The cache returned an error. Log the error and proceed with the rest of this
// function.
Err(e) => warn!(
self.log,
"Early attester cache failed";
"error" => ?e
),
}
let slots_per_epoch = T::EthSpec::slots_per_epoch();
let request_epoch = request_slot.epoch(slots_per_epoch);
/*
* Phase 1/2:
*
* Take a short-lived read-lock on the head and copy the necessary information from it.
*
* It is important that this first phase is as quick as possible; creating contention for
* the head-lock is not desirable.
*/
let head_state_slot;
let beacon_block_root;
let beacon_state_root;
let target;
let current_epoch_attesting_info: Option<(Checkpoint, usize)>;
let attester_cache_key;
let head_timer = metrics::start_timer(&metrics::ATTESTATION_PRODUCTION_HEAD_SCRAPE_SECONDS);
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// The following braces are to prevent the `cached_head` Arc from being held for longer than
// required. It also helps reduce the diff for a very large PR (#3244).
{
let head = self.head_snapshot();
let head_state = &head.beacon_state;
head_state_slot = head_state.slot();
// There is no value in producing an attestation to a block that is pre-finalization and
// it is likely to cause expensive and pointless reads to the freezer database. Exit
// early if this is the case.
let finalized_slot = head_state
.finalized_checkpoint()
.epoch
.start_slot(slots_per_epoch);
if request_slot < finalized_slot {
return Err(Error::AttestingToFinalizedSlot {
finalized_slot,
request_slot,
});
}
// This function will eventually fail when trying to access a slot which is
// out-of-bounds of `state.block_roots`. This explicit error is intended to provide a
// clearer message to the user than an ambiguous `SlotOutOfBounds` error.
let slots_per_historical_root = T::EthSpec::slots_per_historical_root() as u64;
let lowest_permissible_slot =
head_state.slot().saturating_sub(slots_per_historical_root);
if request_slot < lowest_permissible_slot {
return Err(Error::AttestingToAncientSlot {
lowest_permissible_slot,
request_slot,
});
}
if request_slot >= head_state.slot() {
// When attesting to the head slot or later, always use the head of the chain.
beacon_block_root = head.beacon_block_root;
beacon_state_root = head.beacon_state_root();
} else {
// Permit attesting to slots *prior* to the current head. This is desirable when
// the VC and BN are out-of-sync due to time issues or overloading.
beacon_block_root = *head_state.get_block_root(request_slot)?;
beacon_state_root = *head_state.get_state_root(request_slot)?;
};
let target_slot = request_epoch.start_slot(T::EthSpec::slots_per_epoch());
let target_root = if head_state.slot() <= target_slot {
// If the state is earlier than the target slot then the target *must* be the head
// block root.
beacon_block_root
} else {
*head_state.get_block_root(target_slot)?
};
target = Checkpoint {
epoch: request_epoch,
root: target_root,
};
current_epoch_attesting_info = if head_state.current_epoch() == request_epoch {
// When the head state is in the same epoch as the request, all the information
// required to attest is available on the head state.
Some((
head_state.current_justified_checkpoint(),
head_state
.get_beacon_committee(request_slot, request_index)?
.committee
.len(),
))
} else {
// If the head state is in a *different* epoch to the request, more work is required
// to determine the justified checkpoint and committee length.
None
};
// Determine the key for `self.attester_cache`, in case it is required later in this
// routine.
attester_cache_key =
AttesterCacheKey::new(request_epoch, head_state, beacon_block_root)?;
Faster attestation production (#838) * Start adding interop genesis state to lcli * Use more efficient method to generate genesis state * Remove duplicate int_to_bytes32 * Add lcli command to change state genesis time * Add option to allow VC to start with unsynced BN * Set VC to do parallel key loading * Don't default to dummy eth1 backend * Add endpoint to dump operation pool * Add metrics for op pool * Remove state clone for slot notifier * Add mem size approximation for tree hash cache * Avoid cloning tree hash when getting head * Avoid cloning tree hash when getting head * Add working arena-based cached tree hash * Add another benchmark * Add pre-allocation for caches * Make cache nullable * Fix bugs in cache tree hash * Add validator tree hash optimization * Optimize hash_concat * Make hash32_concat return fixed-len array * Fix failing API tests * Add new beacon state cache struct * Add validator-specific cache * Separate list and values arenas * Add parallel validator registry hashing * Remove MultiTreeHashCache * Remove cached tree hash macro * Fix failing tree hash test * Address Michael's comments * Add CachedTreeHash impl for ef tests * Fix messy merge conflict * Optimize attestation production * Add first basic optimizations * Fix SlotOutOfBounds error * Resolved missed merge conflicts * Fix another missed merge conflict * Fix more merge conflict issues * Add `StateSkipConfig` * Fix test compile errors * Add failing test * Fix bug, make tests pass * Add comment * Delete unused function * Replace deleted comment
2020-03-04 06:10:22 +00:00
}
drop(head_timer);
// Only attest to a block if it is fully verified (i.e. not optimistic or invalid).
match self
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
.canonical_head
.fork_choice_read_lock()
.get_block_execution_status(&beacon_block_root)
{
Some(execution_status) if execution_status.is_valid_or_irrelevant() => (),
Some(execution_status) => {
return Err(Error::HeadBlockNotFullyVerified {
beacon_block_root,
execution_status,
})
}
None => return Err(Error::HeadMissingFromForkChoice(beacon_block_root)),
};
/*
* Phase 2/2:
*
* If the justified checkpoint and committee length from the head are suitable for this
* attestation, use them. If not, try the attester cache. If the cache misses, load a state
* from disk and prime the cache with it.
*/
let cache_timer =
metrics::start_timer(&metrics::ATTESTATION_PRODUCTION_CACHE_INTERACTION_SECONDS);
let (justified_checkpoint, committee_len) =
if let Some((justified_checkpoint, committee_len)) = current_epoch_attesting_info {
// The head state is in the same epoch as the attestation, so there is no more
// required information.
(justified_checkpoint, committee_len)
} else if let Some(cached_values) = self.attester_cache.get::<T::EthSpec>(
&attester_cache_key,
request_slot,
request_index,
&self.spec,
)? {
// The suitable values were already cached. Return them.
cached_values
} else {
debug!(
self.log,
"Attester cache miss";
"beacon_block_root" => ?beacon_block_root,
"head_state_slot" => %head_state_slot,
"request_slot" => %request_slot,
);
// Neither the head state, nor the attester cache was able to produce the required
// information to attest in this epoch. So, load a `BeaconState` from disk and use
// it to fulfil the request (and prime the cache to avoid this next time).
let _cache_build_timer =
metrics::start_timer(&metrics::ATTESTATION_PRODUCTION_CACHE_PRIME_SECONDS);
self.attester_cache.load_and_cache_state(
beacon_state_root,
attester_cache_key,
request_slot,
request_index,
self,
)?
};
drop(cache_timer);
Ok(Attestation {
aggregation_bits: BitList::with_capacity(committee_len)?,
data: AttestationData {
slot: request_slot,
index: request_index,
beacon_block_root,
source: justified_checkpoint,
target,
},
signature: AggregateSignature::empty(),
})
2019-06-20 08:46:03 +00:00
}
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
/// Performs the same validation as `Self::verify_unaggregated_attestation_for_gossip`, but for
/// multiple attestations using batch BLS verification. Batch verification can provide
/// significant CPU-time savings compared to individual verification.
pub fn batch_verify_unaggregated_attestations_for_gossip<'a, I>(
&self,
attestations: I,
) -> Result<
Vec<Result<VerifiedUnaggregatedAttestation<'a, T>, AttestationError>>,
AttestationError,
>
where
I: Iterator<Item = (&'a Attestation<T::EthSpec>, Option<SubnetId>)> + ExactSizeIterator,
{
batch_verify_unaggregated_attestations(attestations, self)
}
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
/// Accepts some `Attestation` from the network and attempts to verify it, returning `Ok(_)` if
/// it is valid to be (re)broadcast on the gossip network.
Attestation processing (#497) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Fix minor PR comments * Remove duplicated attestation finalization check * Remove awkward `let` statement
2019-08-14 00:55:24 +00:00
///
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
/// The attestation must be "unaggregated", that is it must have exactly one
/// aggregation bit set.
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
pub fn verify_unaggregated_attestation_for_gossip<'a>(
&self,
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
unaggregated_attestation: &'a Attestation<T::EthSpec>,
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
subnet_id: Option<SubnetId>,
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
) -> Result<VerifiedUnaggregatedAttestation<'a, T>, AttestationError> {
Stable futures (#879) * Port eth1 lib to use stable futures * Port eth1_test_rig to stable futures * Port eth1 tests to stable futures * Port genesis service to stable futures * Port genesis tests to stable futures * Port beacon_chain to stable futures * Port lcli to stable futures * Fix eth1_test_rig (#1014) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Update hashmap hashset to stable futures * Adds panic test to hashset delay * Port remote_beacon_node to stable futures * Fix lcli merge conflicts * Non rpc stuff compiles * protocol.rs compiles * Port websockets, timer and notifier to stable futures (#1035) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Port remote_beacon_node to stable futures * Partial eth2-libp2p stable future upgrade * Finished first round of fighting RPC types * Further progress towards porting eth2-libp2p adds caching to discovery * Update behaviour * RPC handler to stable futures * Update RPC to master libp2p * Network service additions * Fix the fallback transport construction (#1102) * Correct warning * Remove hashmap delay * Compiling version of eth2-libp2p * Update all crates versions * Fix conversion function and add tests (#1113) * Port validator_client to stable futures (#1114) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Revert "Update hashmap hashset to stable futures" This reverts commit d432378a3cc5cd67fc29c0b15b96b886c1323554. * Revert "Adds panic test to hashset delay" This reverts commit 281502396fc5b90d9c421a309c2c056982c9525b. * Ported attestation_service * Ported duties_service * Ported fork_service * More ports * Port block_service * Minor fixes * VC compiles * Update TODOS * Borrow self where possible * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test * Re-Fix attestation and duties service * Minor fixes Co-authored-by: Paul Hauner <paul@paulhauner.com> * Network crate update to stable futures * Port account_manager to stable futures (#1121) * Port account_manager to stable futures * Run async fns in tokio environment * Port rest_api crate to stable futures (#1118) * Port rest_api lib to stable futures * Reduce tokio features * Update notifier to stable futures * Builder update * Further updates * Convert self referential async functions * stable futures fixes (#1124) * Fix eth1 update functions * Fix genesis and client * Fix beacon node lib * Return appropriate runtimes from environment * Fix test rig * Refactor eth1 service update * Upgrade simulator to stable futures * Lighthouse compiles on stable futures * Remove println debugging statement * Update libp2p service, start rpc test upgrade * Update network crate for new libp2p * Update tokio::codec to futures_codec (#1128) * Further work towards RPC corrections * Correct http timeout and network service select * Use tokio runtime for libp2p * Revert "Update tokio::codec to futures_codec (#1128)" This reverts commit e57aea924acf5cbabdcea18895ac07e38a425ed7. * Upgrade RPC libp2p tests * Upgrade secio fallback test * Upgrade gossipsub examples * Clean up RPC protocol * Test fixes (#1133) * Correct websocket timeout and run on os thread * Fix network test * Clean up PR * Correct tokio tcp move attestation service tests * Upgrade attestation service tests * Correct network test * Correct genesis test * Test corrections * Log info when block is received * Modify logs and update attester service events * Stable futures: fixes to vc, eth1 and account manager (#1142) * Add local testnet scripts * Remove whiteblock script * Rename local testnet script * Move spawns onto handle * Fix VC panic * Initial fix to block production issue * Tidy block producer fix * Tidy further * Add local testnet clean script * Run cargo fmt * Tidy duties service * Tidy fork service * Tidy ForkService * Tidy AttestationService * Tidy notifier * Ensure await is not suppressed in eth1 * Ensure await is not suppressed in account_manager * Use .ok() instead of .unwrap_or(()) * RPC decoding test for proto * Update discv5 and eth2-libp2p deps * Fix lcli double runtime issue (#1144) * Handle stream termination and dialing peer errors * Correct peer_info variant types * Remove unnecessary warnings * Handle subnet unsubscription removal and improve logigng * Add logs around ping * Upgrade discv5 and improve logging * Handle peer connection status for multiple connections * Improve network service logging * Improve logging around peer manager * Upgrade swarm poll centralise peer management * Identify clients on error * Fix `remove_peer` in sync (#1150) * remove_peer removes from all chains * Remove logs * Fix early return from loop * Improved logging, fix panic * Partially correct tests * Stable futures: Vc sync (#1149) * Improve syncing heuristic * Add comments * Use safer method for tolerance * Fix tests * Stable futures: Fix VC bug, update agg pool, add more metrics (#1151) * Expose epoch processing summary * Expose participation metrics to prometheus * Switch to f64 * Reduce precision * Change precision * Expose observed attesters metrics * Add metrics for agg/unagg attn counts * Add metrics for gossip rx * Add metrics for gossip tx * Adds ignored attns to prom * Add attestation timing * Add timer for aggregation pool sig agg * Add write lock timer for agg pool * Add more metrics to agg pool * Change map lock code * Add extra metric to agg pool * Change lock handling in agg pool * Change .write() to .read() * Add another agg pool timer * Fix for is_aggregator * Fix pruning bug Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-05-17 11:16:48 +00:00
metrics::inc_counter(&metrics::UNAGGREGATED_ATTESTATION_PROCESSING_REQUESTS);
let _timer =
metrics::start_timer(&metrics::UNAGGREGATED_ATTESTATION_GOSSIP_VERIFICATION_TIMES);
VerifiedUnaggregatedAttestation::verify(unaggregated_attestation, subnet_id, self).map(
|v| {
// This method is called for API and gossip attestations, so this covers all unaggregated attestation events
if let Some(event_handler) = self.event_handler.as_ref() {
if event_handler.has_attestation_subscribers() {
event_handler
.register(EventKind::Attestation(Box::new(v.attestation().clone())));
}
}
metrics::inc_counter(&metrics::UNAGGREGATED_ATTESTATION_PROCESSING_SUCCESSES);
v
},
)
}
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
/// Performs the same validation as `Self::verify_aggregated_attestation_for_gossip`, but for
/// multiple attestations using batch BLS verification. Batch verification can provide
/// significant CPU-time savings compared to individual verification.
pub fn batch_verify_aggregated_attestations_for_gossip<'a, I>(
&self,
aggregates: I,
) -> Result<Vec<Result<VerifiedAggregatedAttestation<'a, T>, AttestationError>>, AttestationError>
where
I: Iterator<Item = &'a SignedAggregateAndProof<T::EthSpec>> + ExactSizeIterator,
{
batch_verify_aggregated_attestations(aggregates, self)
}
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
/// Accepts some `SignedAggregateAndProof` from the network and attempts to verify it,
/// returning `Ok(_)` if it is valid to be (re)broadcast on the gossip network.
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
pub fn verify_aggregated_attestation_for_gossip<'a>(
&self,
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
signed_aggregate: &'a SignedAggregateAndProof<T::EthSpec>,
) -> Result<VerifiedAggregatedAttestation<'a, T>, AttestationError> {
Stable futures (#879) * Port eth1 lib to use stable futures * Port eth1_test_rig to stable futures * Port eth1 tests to stable futures * Port genesis service to stable futures * Port genesis tests to stable futures * Port beacon_chain to stable futures * Port lcli to stable futures * Fix eth1_test_rig (#1014) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Update hashmap hashset to stable futures * Adds panic test to hashset delay * Port remote_beacon_node to stable futures * Fix lcli merge conflicts * Non rpc stuff compiles * protocol.rs compiles * Port websockets, timer and notifier to stable futures (#1035) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Port remote_beacon_node to stable futures * Partial eth2-libp2p stable future upgrade * Finished first round of fighting RPC types * Further progress towards porting eth2-libp2p adds caching to discovery * Update behaviour * RPC handler to stable futures * Update RPC to master libp2p * Network service additions * Fix the fallback transport construction (#1102) * Correct warning * Remove hashmap delay * Compiling version of eth2-libp2p * Update all crates versions * Fix conversion function and add tests (#1113) * Port validator_client to stable futures (#1114) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Revert "Update hashmap hashset to stable futures" This reverts commit d432378a3cc5cd67fc29c0b15b96b886c1323554. * Revert "Adds panic test to hashset delay" This reverts commit 281502396fc5b90d9c421a309c2c056982c9525b. * Ported attestation_service * Ported duties_service * Ported fork_service * More ports * Port block_service * Minor fixes * VC compiles * Update TODOS * Borrow self where possible * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test * Re-Fix attestation and duties service * Minor fixes Co-authored-by: Paul Hauner <paul@paulhauner.com> * Network crate update to stable futures * Port account_manager to stable futures (#1121) * Port account_manager to stable futures * Run async fns in tokio environment * Port rest_api crate to stable futures (#1118) * Port rest_api lib to stable futures * Reduce tokio features * Update notifier to stable futures * Builder update * Further updates * Convert self referential async functions * stable futures fixes (#1124) * Fix eth1 update functions * Fix genesis and client * Fix beacon node lib * Return appropriate runtimes from environment * Fix test rig * Refactor eth1 service update * Upgrade simulator to stable futures * Lighthouse compiles on stable futures * Remove println debugging statement * Update libp2p service, start rpc test upgrade * Update network crate for new libp2p * Update tokio::codec to futures_codec (#1128) * Further work towards RPC corrections * Correct http timeout and network service select * Use tokio runtime for libp2p * Revert "Update tokio::codec to futures_codec (#1128)" This reverts commit e57aea924acf5cbabdcea18895ac07e38a425ed7. * Upgrade RPC libp2p tests * Upgrade secio fallback test * Upgrade gossipsub examples * Clean up RPC protocol * Test fixes (#1133) * Correct websocket timeout and run on os thread * Fix network test * Clean up PR * Correct tokio tcp move attestation service tests * Upgrade attestation service tests * Correct network test * Correct genesis test * Test corrections * Log info when block is received * Modify logs and update attester service events * Stable futures: fixes to vc, eth1 and account manager (#1142) * Add local testnet scripts * Remove whiteblock script * Rename local testnet script * Move spawns onto handle * Fix VC panic * Initial fix to block production issue * Tidy block producer fix * Tidy further * Add local testnet clean script * Run cargo fmt * Tidy duties service * Tidy fork service * Tidy ForkService * Tidy AttestationService * Tidy notifier * Ensure await is not suppressed in eth1 * Ensure await is not suppressed in account_manager * Use .ok() instead of .unwrap_or(()) * RPC decoding test for proto * Update discv5 and eth2-libp2p deps * Fix lcli double runtime issue (#1144) * Handle stream termination and dialing peer errors * Correct peer_info variant types * Remove unnecessary warnings * Handle subnet unsubscription removal and improve logigng * Add logs around ping * Upgrade discv5 and improve logging * Handle peer connection status for multiple connections * Improve network service logging * Improve logging around peer manager * Upgrade swarm poll centralise peer management * Identify clients on error * Fix `remove_peer` in sync (#1150) * remove_peer removes from all chains * Remove logs * Fix early return from loop * Improved logging, fix panic * Partially correct tests * Stable futures: Vc sync (#1149) * Improve syncing heuristic * Add comments * Use safer method for tolerance * Fix tests * Stable futures: Fix VC bug, update agg pool, add more metrics (#1151) * Expose epoch processing summary * Expose participation metrics to prometheus * Switch to f64 * Reduce precision * Change precision * Expose observed attesters metrics * Add metrics for agg/unagg attn counts * Add metrics for gossip rx * Add metrics for gossip tx * Adds ignored attns to prom * Add attestation timing * Add timer for aggregation pool sig agg * Add write lock timer for agg pool * Add more metrics to agg pool * Change map lock code * Add extra metric to agg pool * Change lock handling in agg pool * Change .write() to .read() * Add another agg pool timer * Fix for is_aggregator * Fix pruning bug Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-05-17 11:16:48 +00:00
metrics::inc_counter(&metrics::AGGREGATED_ATTESTATION_PROCESSING_REQUESTS);
let _timer =
metrics::start_timer(&metrics::AGGREGATED_ATTESTATION_GOSSIP_VERIFICATION_TIMES);
VerifiedAggregatedAttestation::verify(signed_aggregate, self).map(|v| {
// This method is called for API and gossip attestations, so this covers all aggregated attestation events
if let Some(event_handler) = self.event_handler.as_ref() {
if event_handler.has_attestation_subscribers() {
event_handler
.register(EventKind::Attestation(Box::new(v.attestation().clone())));
}
}
Stable futures (#879) * Port eth1 lib to use stable futures * Port eth1_test_rig to stable futures * Port eth1 tests to stable futures * Port genesis service to stable futures * Port genesis tests to stable futures * Port beacon_chain to stable futures * Port lcli to stable futures * Fix eth1_test_rig (#1014) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Update hashmap hashset to stable futures * Adds panic test to hashset delay * Port remote_beacon_node to stable futures * Fix lcli merge conflicts * Non rpc stuff compiles * protocol.rs compiles * Port websockets, timer and notifier to stable futures (#1035) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Port remote_beacon_node to stable futures * Partial eth2-libp2p stable future upgrade * Finished first round of fighting RPC types * Further progress towards porting eth2-libp2p adds caching to discovery * Update behaviour * RPC handler to stable futures * Update RPC to master libp2p * Network service additions * Fix the fallback transport construction (#1102) * Correct warning * Remove hashmap delay * Compiling version of eth2-libp2p * Update all crates versions * Fix conversion function and add tests (#1113) * Port validator_client to stable futures (#1114) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Revert "Update hashmap hashset to stable futures" This reverts commit d432378a3cc5cd67fc29c0b15b96b886c1323554. * Revert "Adds panic test to hashset delay" This reverts commit 281502396fc5b90d9c421a309c2c056982c9525b. * Ported attestation_service * Ported duties_service * Ported fork_service * More ports * Port block_service * Minor fixes * VC compiles * Update TODOS * Borrow self where possible * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test * Re-Fix attestation and duties service * Minor fixes Co-authored-by: Paul Hauner <paul@paulhauner.com> * Network crate update to stable futures * Port account_manager to stable futures (#1121) * Port account_manager to stable futures * Run async fns in tokio environment * Port rest_api crate to stable futures (#1118) * Port rest_api lib to stable futures * Reduce tokio features * Update notifier to stable futures * Builder update * Further updates * Convert self referential async functions * stable futures fixes (#1124) * Fix eth1 update functions * Fix genesis and client * Fix beacon node lib * Return appropriate runtimes from environment * Fix test rig * Refactor eth1 service update * Upgrade simulator to stable futures * Lighthouse compiles on stable futures * Remove println debugging statement * Update libp2p service, start rpc test upgrade * Update network crate for new libp2p * Update tokio::codec to futures_codec (#1128) * Further work towards RPC corrections * Correct http timeout and network service select * Use tokio runtime for libp2p * Revert "Update tokio::codec to futures_codec (#1128)" This reverts commit e57aea924acf5cbabdcea18895ac07e38a425ed7. * Upgrade RPC libp2p tests * Upgrade secio fallback test * Upgrade gossipsub examples * Clean up RPC protocol * Test fixes (#1133) * Correct websocket timeout and run on os thread * Fix network test * Clean up PR * Correct tokio tcp move attestation service tests * Upgrade attestation service tests * Correct network test * Correct genesis test * Test corrections * Log info when block is received * Modify logs and update attester service events * Stable futures: fixes to vc, eth1 and account manager (#1142) * Add local testnet scripts * Remove whiteblock script * Rename local testnet script * Move spawns onto handle * Fix VC panic * Initial fix to block production issue * Tidy block producer fix * Tidy further * Add local testnet clean script * Run cargo fmt * Tidy duties service * Tidy fork service * Tidy ForkService * Tidy AttestationService * Tidy notifier * Ensure await is not suppressed in eth1 * Ensure await is not suppressed in account_manager * Use .ok() instead of .unwrap_or(()) * RPC decoding test for proto * Update discv5 and eth2-libp2p deps * Fix lcli double runtime issue (#1144) * Handle stream termination and dialing peer errors * Correct peer_info variant types * Remove unnecessary warnings * Handle subnet unsubscription removal and improve logigng * Add logs around ping * Upgrade discv5 and improve logging * Handle peer connection status for multiple connections * Improve network service logging * Improve logging around peer manager * Upgrade swarm poll centralise peer management * Identify clients on error * Fix `remove_peer` in sync (#1150) * remove_peer removes from all chains * Remove logs * Fix early return from loop * Improved logging, fix panic * Partially correct tests * Stable futures: Vc sync (#1149) * Improve syncing heuristic * Add comments * Use safer method for tolerance * Fix tests * Stable futures: Fix VC bug, update agg pool, add more metrics (#1151) * Expose epoch processing summary * Expose participation metrics to prometheus * Switch to f64 * Reduce precision * Change precision * Expose observed attesters metrics * Add metrics for agg/unagg attn counts * Add metrics for gossip rx * Add metrics for gossip tx * Adds ignored attns to prom * Add attestation timing * Add timer for aggregation pool sig agg * Add write lock timer for agg pool * Add more metrics to agg pool * Change map lock code * Add extra metric to agg pool * Change lock handling in agg pool * Change .write() to .read() * Add another agg pool timer * Fix for is_aggregator * Fix pruning bug Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-05-17 11:16:48 +00:00
metrics::inc_counter(&metrics::AGGREGATED_ATTESTATION_PROCESSING_SUCCESSES);
v
})
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
}
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
/// Accepts some `SyncCommitteeMessage` from the network and attempts to verify it, returning `Ok(_)` if
/// it is valid to be (re)broadcast on the gossip network.
pub fn verify_sync_committee_message_for_gossip(
&self,
sync_message: SyncCommitteeMessage,
subnet_id: SyncSubnetId,
) -> Result<VerifiedSyncCommitteeMessage, SyncCommitteeError> {
metrics::inc_counter(&metrics::SYNC_MESSAGE_PROCESSING_REQUESTS);
let _timer = metrics::start_timer(&metrics::SYNC_MESSAGE_GOSSIP_VERIFICATION_TIMES);
VerifiedSyncCommitteeMessage::verify(sync_message, subnet_id, self).map(|v| {
metrics::inc_counter(&metrics::SYNC_MESSAGE_PROCESSING_SUCCESSES);
v
})
}
/// Accepts some `SignedContributionAndProof` from the network and attempts to verify it,
/// returning `Ok(_)` if it is valid to be (re)broadcast on the gossip network.
pub fn verify_sync_contribution_for_gossip(
&self,
sync_contribution: SignedContributionAndProof<T::EthSpec>,
) -> Result<VerifiedSyncContribution<T>, SyncCommitteeError> {
metrics::inc_counter(&metrics::SYNC_CONTRIBUTION_PROCESSING_REQUESTS);
let _timer = metrics::start_timer(&metrics::SYNC_CONTRIBUTION_GOSSIP_VERIFICATION_TIMES);
VerifiedSyncContribution::verify(sync_contribution, self).map(|v| {
if let Some(event_handler) = self.event_handler.as_ref() {
if event_handler.has_contribution_subscribers() {
event_handler.register(EventKind::ContributionAndProof(Box::new(
v.aggregate().clone(),
)));
}
}
metrics::inc_counter(&metrics::SYNC_CONTRIBUTION_PROCESSING_SUCCESSES);
v
})
}
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
/// Accepts some attestation-type object and attempts to verify it in the context of fork
/// choice. If it is valid it is applied to `self.fork_choice`.
///
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
/// Common items that implement `VerifiedAttestation`:
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
///
/// - `VerifiedUnaggregatedAttestation`
/// - `VerifiedAggregatedAttestation`
pub fn apply_attestation_to_fork_choice(
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
&self,
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
verified: &impl VerifiedAttestation<T>,
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
) -> Result<(), Error> {
let _timer = metrics::start_timer(&metrics::FORK_CHOICE_PROCESS_ATTESTATION_TIMES);
Stable futures (#879) * Port eth1 lib to use stable futures * Port eth1_test_rig to stable futures * Port eth1 tests to stable futures * Port genesis service to stable futures * Port genesis tests to stable futures * Port beacon_chain to stable futures * Port lcli to stable futures * Fix eth1_test_rig (#1014) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Update hashmap hashset to stable futures * Adds panic test to hashset delay * Port remote_beacon_node to stable futures * Fix lcli merge conflicts * Non rpc stuff compiles * protocol.rs compiles * Port websockets, timer and notifier to stable futures (#1035) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Port remote_beacon_node to stable futures * Partial eth2-libp2p stable future upgrade * Finished first round of fighting RPC types * Further progress towards porting eth2-libp2p adds caching to discovery * Update behaviour * RPC handler to stable futures * Update RPC to master libp2p * Network service additions * Fix the fallback transport construction (#1102) * Correct warning * Remove hashmap delay * Compiling version of eth2-libp2p * Update all crates versions * Fix conversion function and add tests (#1113) * Port validator_client to stable futures (#1114) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Revert "Update hashmap hashset to stable futures" This reverts commit d432378a3cc5cd67fc29c0b15b96b886c1323554. * Revert "Adds panic test to hashset delay" This reverts commit 281502396fc5b90d9c421a309c2c056982c9525b. * Ported attestation_service * Ported duties_service * Ported fork_service * More ports * Port block_service * Minor fixes * VC compiles * Update TODOS * Borrow self where possible * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test * Re-Fix attestation and duties service * Minor fixes Co-authored-by: Paul Hauner <paul@paulhauner.com> * Network crate update to stable futures * Port account_manager to stable futures (#1121) * Port account_manager to stable futures * Run async fns in tokio environment * Port rest_api crate to stable futures (#1118) * Port rest_api lib to stable futures * Reduce tokio features * Update notifier to stable futures * Builder update * Further updates * Convert self referential async functions * stable futures fixes (#1124) * Fix eth1 update functions * Fix genesis and client * Fix beacon node lib * Return appropriate runtimes from environment * Fix test rig * Refactor eth1 service update * Upgrade simulator to stable futures * Lighthouse compiles on stable futures * Remove println debugging statement * Update libp2p service, start rpc test upgrade * Update network crate for new libp2p * Update tokio::codec to futures_codec (#1128) * Further work towards RPC corrections * Correct http timeout and network service select * Use tokio runtime for libp2p * Revert "Update tokio::codec to futures_codec (#1128)" This reverts commit e57aea924acf5cbabdcea18895ac07e38a425ed7. * Upgrade RPC libp2p tests * Upgrade secio fallback test * Upgrade gossipsub examples * Clean up RPC protocol * Test fixes (#1133) * Correct websocket timeout and run on os thread * Fix network test * Clean up PR * Correct tokio tcp move attestation service tests * Upgrade attestation service tests * Correct network test * Correct genesis test * Test corrections * Log info when block is received * Modify logs and update attester service events * Stable futures: fixes to vc, eth1 and account manager (#1142) * Add local testnet scripts * Remove whiteblock script * Rename local testnet script * Move spawns onto handle * Fix VC panic * Initial fix to block production issue * Tidy block producer fix * Tidy further * Add local testnet clean script * Run cargo fmt * Tidy duties service * Tidy fork service * Tidy ForkService * Tidy AttestationService * Tidy notifier * Ensure await is not suppressed in eth1 * Ensure await is not suppressed in account_manager * Use .ok() instead of .unwrap_or(()) * RPC decoding test for proto * Update discv5 and eth2-libp2p deps * Fix lcli double runtime issue (#1144) * Handle stream termination and dialing peer errors * Correct peer_info variant types * Remove unnecessary warnings * Handle subnet unsubscription removal and improve logigng * Add logs around ping * Upgrade discv5 and improve logging * Handle peer connection status for multiple connections * Improve network service logging * Improve logging around peer manager * Upgrade swarm poll centralise peer management * Identify clients on error * Fix `remove_peer` in sync (#1150) * remove_peer removes from all chains * Remove logs * Fix early return from loop * Improved logging, fix panic * Partially correct tests * Stable futures: Vc sync (#1149) * Improve syncing heuristic * Add comments * Use safer method for tolerance * Fix tests * Stable futures: Fix VC bug, update agg pool, add more metrics (#1151) * Expose epoch processing summary * Expose participation metrics to prometheus * Switch to f64 * Reduce precision * Change precision * Expose observed attesters metrics * Add metrics for agg/unagg attn counts * Add metrics for gossip rx * Add metrics for gossip tx * Adds ignored attns to prom * Add attestation timing * Add timer for aggregation pool sig agg * Add write lock timer for agg pool * Add more metrics to agg pool * Change map lock code * Add extra metric to agg pool * Change lock handling in agg pool * Change .write() to .read() * Add another agg pool timer * Fix for is_aggregator * Fix pruning bug Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-05-17 11:16:48 +00:00
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
self.canonical_head
.fork_choice_write_lock()
v1.1.6 Fork Choice changes (#2822) ## Issue Addressed Resolves: https://github.com/sigp/lighthouse/issues/2741 Includes: https://github.com/sigp/lighthouse/pull/2853 so that we can get ssz static tests passing here on v1.1.6. If we want to merge that first, we can make this diff slightly smaller ## Proposed Changes - Changes the `justified_epoch` and `finalized_epoch` in the `ProtoArrayNode` each to an `Option<Checkpoint>`. The `Option` is necessary only for the migration, so not ideal. But does allow us to add a default logic to `None` on these fields during the database migration. - Adds a database migration from a legacy fork choice struct to the new one, search for all necessary block roots in fork choice by iterating through blocks in the db. - updates related to https://github.com/ethereum/consensus-specs/pull/2727 - We will have to update the persisted forkchoice to make sure the justified checkpoint stored is correct according to the updated fork choice logic. This boils down to setting the forkchoice store's justified checkpoint to the justified checkpoint of the block that advanced the finalized checkpoint to the current one. - AFAICT there's no migration steps necessary for the update to allow applying attestations from prior blocks, but would appreciate confirmation on that - I updated the consensus spec tests to v1.1.6 here, but they will fail until we also implement the proposer score boost updates. I confirmed that the previously failing scenario `new_finalized_slot_is_justified_checkpoint_ancestor` will now pass after the boost updates, but haven't confirmed _all_ tests will pass because I just quickly stubbed out the proposer boost test scenario formatting. - This PR now also includes proposer boosting https://github.com/ethereum/consensus-specs/pull/2730 ## Additional Info I realized checking justified and finalized roots in fork choice makes it more likely that we trigger this bug: https://github.com/ethereum/consensus-specs/pull/2727 It's possible the combination of justified checkpoint and finalized checkpoint in the forkchoice store is different from in any block in fork choice. So when trying to startup our store's justified checkpoint seems invalid to the rest of fork choice (but it should be valid). When this happens we get an `InvalidBestNode` error and fail to start up. So I'm including that bugfix in this branch. Todo: - [x] Fix fork choice tests - [x] Self review - [x] Add fix for https://github.com/ethereum/consensus-specs/pull/2727 - [x] Rebase onto Kintusgi - [x] Fix `num_active_validators` calculation as @michaelsproul pointed out - [x] Clean up db migrations Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-12-13 20:43:22 +00:00
.on_attestation(
self.slot()?,
verified.indexed_attestation(),
AttestationFromBlock::False,
&self.spec,
v1.1.6 Fork Choice changes (#2822) ## Issue Addressed Resolves: https://github.com/sigp/lighthouse/issues/2741 Includes: https://github.com/sigp/lighthouse/pull/2853 so that we can get ssz static tests passing here on v1.1.6. If we want to merge that first, we can make this diff slightly smaller ## Proposed Changes - Changes the `justified_epoch` and `finalized_epoch` in the `ProtoArrayNode` each to an `Option<Checkpoint>`. The `Option` is necessary only for the migration, so not ideal. But does allow us to add a default logic to `None` on these fields during the database migration. - Adds a database migration from a legacy fork choice struct to the new one, search for all necessary block roots in fork choice by iterating through blocks in the db. - updates related to https://github.com/ethereum/consensus-specs/pull/2727 - We will have to update the persisted forkchoice to make sure the justified checkpoint stored is correct according to the updated fork choice logic. This boils down to setting the forkchoice store's justified checkpoint to the justified checkpoint of the block that advanced the finalized checkpoint to the current one. - AFAICT there's no migration steps necessary for the update to allow applying attestations from prior blocks, but would appreciate confirmation on that - I updated the consensus spec tests to v1.1.6 here, but they will fail until we also implement the proposer score boost updates. I confirmed that the previously failing scenario `new_finalized_slot_is_justified_checkpoint_ancestor` will now pass after the boost updates, but haven't confirmed _all_ tests will pass because I just quickly stubbed out the proposer boost test scenario formatting. - This PR now also includes proposer boosting https://github.com/ethereum/consensus-specs/pull/2730 ## Additional Info I realized checking justified and finalized roots in fork choice makes it more likely that we trigger this bug: https://github.com/ethereum/consensus-specs/pull/2727 It's possible the combination of justified checkpoint and finalized checkpoint in the forkchoice store is different from in any block in fork choice. So when trying to startup our store's justified checkpoint seems invalid to the rest of fork choice (but it should be valid). When this happens we get an `InvalidBestNode` error and fail to start up. So I'm including that bugfix in this branch. Todo: - [x] Fix fork choice tests - [x] Self review - [x] Add fix for https://github.com/ethereum/consensus-specs/pull/2727 - [x] Rebase onto Kintusgi - [x] Fix `num_active_validators` calculation as @michaelsproul pointed out - [x] Clean up db migrations Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-12-13 20:43:22 +00:00
)
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
.map_err(Into::into)
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
}
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
/// Accepts an `VerifiedUnaggregatedAttestation` and attempts to apply it to the "naive
/// aggregation pool".
///
/// The naive aggregation pool is used by local validators to produce
/// `SignedAggregateAndProof`.
///
/// If the attestation is too old (low slot) to be included in the pool it is simply dropped
/// and no error is returned.
pub fn add_to_naive_aggregation_pool(
&self,
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
unaggregated_attestation: &impl VerifiedAttestation<T>,
) -> Result<(), AttestationError> {
Stable futures (#879) * Port eth1 lib to use stable futures * Port eth1_test_rig to stable futures * Port eth1 tests to stable futures * Port genesis service to stable futures * Port genesis tests to stable futures * Port beacon_chain to stable futures * Port lcli to stable futures * Fix eth1_test_rig (#1014) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Update hashmap hashset to stable futures * Adds panic test to hashset delay * Port remote_beacon_node to stable futures * Fix lcli merge conflicts * Non rpc stuff compiles * protocol.rs compiles * Port websockets, timer and notifier to stable futures (#1035) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Port remote_beacon_node to stable futures * Partial eth2-libp2p stable future upgrade * Finished first round of fighting RPC types * Further progress towards porting eth2-libp2p adds caching to discovery * Update behaviour * RPC handler to stable futures * Update RPC to master libp2p * Network service additions * Fix the fallback transport construction (#1102) * Correct warning * Remove hashmap delay * Compiling version of eth2-libp2p * Update all crates versions * Fix conversion function and add tests (#1113) * Port validator_client to stable futures (#1114) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Revert "Update hashmap hashset to stable futures" This reverts commit d432378a3cc5cd67fc29c0b15b96b886c1323554. * Revert "Adds panic test to hashset delay" This reverts commit 281502396fc5b90d9c421a309c2c056982c9525b. * Ported attestation_service * Ported duties_service * Ported fork_service * More ports * Port block_service * Minor fixes * VC compiles * Update TODOS * Borrow self where possible * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test * Re-Fix attestation and duties service * Minor fixes Co-authored-by: Paul Hauner <paul@paulhauner.com> * Network crate update to stable futures * Port account_manager to stable futures (#1121) * Port account_manager to stable futures * Run async fns in tokio environment * Port rest_api crate to stable futures (#1118) * Port rest_api lib to stable futures * Reduce tokio features * Update notifier to stable futures * Builder update * Further updates * Convert self referential async functions * stable futures fixes (#1124) * Fix eth1 update functions * Fix genesis and client * Fix beacon node lib * Return appropriate runtimes from environment * Fix test rig * Refactor eth1 service update * Upgrade simulator to stable futures * Lighthouse compiles on stable futures * Remove println debugging statement * Update libp2p service, start rpc test upgrade * Update network crate for new libp2p * Update tokio::codec to futures_codec (#1128) * Further work towards RPC corrections * Correct http timeout and network service select * Use tokio runtime for libp2p * Revert "Update tokio::codec to futures_codec (#1128)" This reverts commit e57aea924acf5cbabdcea18895ac07e38a425ed7. * Upgrade RPC libp2p tests * Upgrade secio fallback test * Upgrade gossipsub examples * Clean up RPC protocol * Test fixes (#1133) * Correct websocket timeout and run on os thread * Fix network test * Clean up PR * Correct tokio tcp move attestation service tests * Upgrade attestation service tests * Correct network test * Correct genesis test * Test corrections * Log info when block is received * Modify logs and update attester service events * Stable futures: fixes to vc, eth1 and account manager (#1142) * Add local testnet scripts * Remove whiteblock script * Rename local testnet script * Move spawns onto handle * Fix VC panic * Initial fix to block production issue * Tidy block producer fix * Tidy further * Add local testnet clean script * Run cargo fmt * Tidy duties service * Tidy fork service * Tidy ForkService * Tidy AttestationService * Tidy notifier * Ensure await is not suppressed in eth1 * Ensure await is not suppressed in account_manager * Use .ok() instead of .unwrap_or(()) * RPC decoding test for proto * Update discv5 and eth2-libp2p deps * Fix lcli double runtime issue (#1144) * Handle stream termination and dialing peer errors * Correct peer_info variant types * Remove unnecessary warnings * Handle subnet unsubscription removal and improve logigng * Add logs around ping * Upgrade discv5 and improve logging * Handle peer connection status for multiple connections * Improve network service logging * Improve logging around peer manager * Upgrade swarm poll centralise peer management * Identify clients on error * Fix `remove_peer` in sync (#1150) * remove_peer removes from all chains * Remove logs * Fix early return from loop * Improved logging, fix panic * Partially correct tests * Stable futures: Vc sync (#1149) * Improve syncing heuristic * Add comments * Use safer method for tolerance * Fix tests * Stable futures: Fix VC bug, update agg pool, add more metrics (#1151) * Expose epoch processing summary * Expose participation metrics to prometheus * Switch to f64 * Reduce precision * Change precision * Expose observed attesters metrics * Add metrics for agg/unagg attn counts * Add metrics for gossip rx * Add metrics for gossip tx * Adds ignored attns to prom * Add attestation timing * Add timer for aggregation pool sig agg * Add write lock timer for agg pool * Add more metrics to agg pool * Change map lock code * Add extra metric to agg pool * Change lock handling in agg pool * Change .write() to .read() * Add another agg pool timer * Fix for is_aggregator * Fix pruning bug Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-05-17 11:16:48 +00:00
let _timer = metrics::start_timer(&metrics::ATTESTATION_PROCESSING_APPLY_TO_AGG_POOL);
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
let attestation = unaggregated_attestation.attestation();
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
match self.naive_aggregation_pool.write().insert(attestation) {
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
Ok(outcome) => trace!(
self.log,
"Stored unaggregated attestation";
"outcome" => ?outcome,
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
"index" => attestation.data.index,
"slot" => attestation.data.slot.as_u64(),
),
Err(NaiveAggregationError::SlotTooLow {
slot,
lowest_permissible_slot,
}) => {
trace!(
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
self.log,
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
"Refused to store unaggregated attestation";
"lowest_permissible_slot" => lowest_permissible_slot.as_u64(),
"slot" => slot.as_u64(),
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
);
}
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
Err(e) => {
error!(
self.log,
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
"Failed to store unaggregated attestation";
"error" => ?e,
"index" => attestation.data.index,
"slot" => attestation.data.slot.as_u64(),
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
);
return Err(Error::from(e).into());
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
}
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
};
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
Ok(())
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
}
Attestation processing (#497) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Fix minor PR comments * Remove duplicated attestation finalization check * Remove awkward `let` statement
2019-08-14 00:55:24 +00:00
/// Accepts a `VerifiedSyncCommitteeMessage` and attempts to apply it to the "naive
/// aggregation pool".
///
/// The naive aggregation pool is used by local validators to produce
/// `SignedContributionAndProof`.
///
/// If the sync message is too old (low slot) to be included in the pool it is simply dropped
/// and no error is returned.
pub fn add_to_naive_sync_aggregation_pool(
&self,
verified_sync_committee_message: VerifiedSyncCommitteeMessage,
) -> Result<VerifiedSyncCommitteeMessage, SyncCommitteeError> {
let sync_message = verified_sync_committee_message.sync_message();
let positions_by_subnet_id: &HashMap<SyncSubnetId, Vec<usize>> =
verified_sync_committee_message.subnet_positions();
for (subnet_id, positions) in positions_by_subnet_id.iter() {
for position in positions {
let _timer =
metrics::start_timer(&metrics::SYNC_CONTRIBUTION_PROCESSING_APPLY_TO_AGG_POOL);
let contribution = SyncCommitteeContribution::from_message(
sync_message,
subnet_id.into(),
*position,
)?;
match self
.naive_sync_aggregation_pool
.write()
.insert(&contribution)
{
Ok(outcome) => trace!(
self.log,
"Stored unaggregated sync committee message";
"outcome" => ?outcome,
"index" => sync_message.validator_index,
"slot" => sync_message.slot.as_u64(),
),
Err(NaiveAggregationError::SlotTooLow {
slot,
lowest_permissible_slot,
}) => {
trace!(
self.log,
"Refused to store unaggregated sync committee message";
"lowest_permissible_slot" => lowest_permissible_slot.as_u64(),
"slot" => slot.as_u64(),
);
}
Err(e) => {
error!(
self.log,
"Failed to store unaggregated sync committee message";
"error" => ?e,
"index" => sync_message.validator_index,
"slot" => sync_message.slot.as_u64(),
);
return Err(Error::from(e).into());
}
};
}
}
Ok(verified_sync_committee_message)
}
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
/// Accepts a `VerifiedAttestation` and attempts to apply it to `self.op_pool`.
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
///
/// The op pool is used by local block producers to pack blocks with operations.
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
pub fn add_to_block_inclusion_pool<A>(
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
&self,
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
verified_attestation: A,
) -> Result<(), AttestationError>
where
A: VerifiedAttestation<T>,
{
Stable futures (#879) * Port eth1 lib to use stable futures * Port eth1_test_rig to stable futures * Port eth1 tests to stable futures * Port genesis service to stable futures * Port genesis tests to stable futures * Port beacon_chain to stable futures * Port lcli to stable futures * Fix eth1_test_rig (#1014) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Update hashmap hashset to stable futures * Adds panic test to hashset delay * Port remote_beacon_node to stable futures * Fix lcli merge conflicts * Non rpc stuff compiles * protocol.rs compiles * Port websockets, timer and notifier to stable futures (#1035) * Fix lcli * Port timer to stable futures * Fix timer * Port websocket_server to stable futures * Port notifier to stable futures * Add TODOS * Port remote_beacon_node to stable futures * Partial eth2-libp2p stable future upgrade * Finished first round of fighting RPC types * Further progress towards porting eth2-libp2p adds caching to discovery * Update behaviour * RPC handler to stable futures * Update RPC to master libp2p * Network service additions * Fix the fallback transport construction (#1102) * Correct warning * Remove hashmap delay * Compiling version of eth2-libp2p * Update all crates versions * Fix conversion function and add tests (#1113) * Port validator_client to stable futures (#1114) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Revert "Update hashmap hashset to stable futures" This reverts commit d432378a3cc5cd67fc29c0b15b96b886c1323554. * Revert "Adds panic test to hashset delay" This reverts commit 281502396fc5b90d9c421a309c2c056982c9525b. * Ported attestation_service * Ported duties_service * Ported fork_service * More ports * Port block_service * Minor fixes * VC compiles * Update TODOS * Borrow self where possible * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test * Re-Fix attestation and duties service * Minor fixes Co-authored-by: Paul Hauner <paul@paulhauner.com> * Network crate update to stable futures * Port account_manager to stable futures (#1121) * Port account_manager to stable futures * Run async fns in tokio environment * Port rest_api crate to stable futures (#1118) * Port rest_api lib to stable futures * Reduce tokio features * Update notifier to stable futures * Builder update * Further updates * Convert self referential async functions * stable futures fixes (#1124) * Fix eth1 update functions * Fix genesis and client * Fix beacon node lib * Return appropriate runtimes from environment * Fix test rig * Refactor eth1 service update * Upgrade simulator to stable futures * Lighthouse compiles on stable futures * Remove println debugging statement * Update libp2p service, start rpc test upgrade * Update network crate for new libp2p * Update tokio::codec to futures_codec (#1128) * Further work towards RPC corrections * Correct http timeout and network service select * Use tokio runtime for libp2p * Revert "Update tokio::codec to futures_codec (#1128)" This reverts commit e57aea924acf5cbabdcea18895ac07e38a425ed7. * Upgrade RPC libp2p tests * Upgrade secio fallback test * Upgrade gossipsub examples * Clean up RPC protocol * Test fixes (#1133) * Correct websocket timeout and run on os thread * Fix network test * Clean up PR * Correct tokio tcp move attestation service tests * Upgrade attestation service tests * Correct network test * Correct genesis test * Test corrections * Log info when block is received * Modify logs and update attester service events * Stable futures: fixes to vc, eth1 and account manager (#1142) * Add local testnet scripts * Remove whiteblock script * Rename local testnet script * Move spawns onto handle * Fix VC panic * Initial fix to block production issue * Tidy block producer fix * Tidy further * Add local testnet clean script * Run cargo fmt * Tidy duties service * Tidy fork service * Tidy ForkService * Tidy AttestationService * Tidy notifier * Ensure await is not suppressed in eth1 * Ensure await is not suppressed in account_manager * Use .ok() instead of .unwrap_or(()) * RPC decoding test for proto * Update discv5 and eth2-libp2p deps * Fix lcli double runtime issue (#1144) * Handle stream termination and dialing peer errors * Correct peer_info variant types * Remove unnecessary warnings * Handle subnet unsubscription removal and improve logigng * Add logs around ping * Upgrade discv5 and improve logging * Handle peer connection status for multiple connections * Improve network service logging * Improve logging around peer manager * Upgrade swarm poll centralise peer management * Identify clients on error * Fix `remove_peer` in sync (#1150) * remove_peer removes from all chains * Remove logs * Fix early return from loop * Improved logging, fix panic * Partially correct tests * Stable futures: Vc sync (#1149) * Improve syncing heuristic * Add comments * Use safer method for tolerance * Fix tests * Stable futures: Fix VC bug, update agg pool, add more metrics (#1151) * Expose epoch processing summary * Expose participation metrics to prometheus * Switch to f64 * Reduce precision * Change precision * Expose observed attesters metrics * Add metrics for agg/unagg attn counts * Add metrics for gossip rx * Add metrics for gossip tx * Adds ignored attns to prom * Add attestation timing * Add timer for aggregation pool sig agg * Add write lock timer for agg pool * Add more metrics to agg pool * Change map lock code * Add extra metric to agg pool * Change lock handling in agg pool * Change .write() to .read() * Add another agg pool timer * Fix for is_aggregator * Fix pruning bug Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-05-17 11:16:48 +00:00
let _timer = metrics::start_timer(&metrics::ATTESTATION_PROCESSING_APPLY_TO_OP_POOL);
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
// If there's no eth1 chain then it's impossible to produce blocks and therefore
// useless to put things in the op pool.
if self.eth1_chain.is_some() {
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
let (attestation, attesting_indices) =
verified_attestation.into_attestation_and_indices();
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
self.op_pool
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
.insert_attestation(attestation, attesting_indices)
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
.map_err(Error::from)?;
Fix block processing blowup, upgrade metrics (#500) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Begin metrics refactor * Move beacon_chain to new metrics structure. * Make metrics not panic if already defined * Use global prometheus gather at rest api * Unify common metric fns into a crate * Add heavy metering to block processing * Remove hypen from prometheus metric name * Add more beacon chain metrics * Add beacon chain persistence metric * Prune op pool on finalization * Add extra prom beacon chain metrics * Prefix BeaconChain metrics with "beacon_" * Add more store metrics * Add basic metrics to libp2p * Add metrics to HTTP server * Remove old `http_server` crate * Update metrics names to be more like standard * Fix broken beacon chain metrics, add slot clock metrics * Add lighthouse_metrics gather fn * Remove http args * Fix wrong state given to op pool prune * Make prom metric names more consistent * Add more metrics, tidy existing metrics * Fix store block read metrics * Tidy attestation metrics * Fix minor PR comments * Allow travis failures on beta (see desc) There's a non-backward compatible change in `cargo fmt`. Stable and beta do not agree. * Tidy `lighthouse_metrics` docs * Fix typo
2019-08-19 11:02:34 +00:00
}
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
Batch BLS verification for attestations (#2399) ## Issue Addressed NA ## Proposed Changes Adds the ability to verify batches of aggregated/unaggregated attestations from the network. When the `BeaconProcessor` finds there are messages in the aggregated or unaggregated attestation queues, it will first check the length of the queue: - `== 1` verify the attestation individually. - `>= 2` take up to 64 of those attestations and verify them in a batch. Notably, we only perform batch verification if the queue has a backlog. We don't apply any artificial delays to attestations to try and force them into batches. ### Batching Details To assist with implementing batches we modify `beacon_chain::attestation_verification` to have two distinct categories for attestations: - *Indexed* attestations: those which have passed initial validation and were valid enough for us to derive an `IndexedAttestation`. - *Verified* attestations: those attestations which were indexed *and also* passed signature verification. These are well-formed, interesting messages which were signed by validators. The batching functions accept `n` attestations and then return `n` attestation verification `Result`s, where those `Result`s can be any combination of `Ok` or `Err`. In other words, we attempt to verify as many attestations as possible and return specific per-attestation results so peer scores can be updated, if required. When we batch verify attestations, we first try to map all those attestations to *indexed* attestations. If any of those attestations were able to be indexed, we then perform batch BLS verification on those indexed attestations. If the batch verification succeeds, we convert them into *verified* attestations, disabling individual signature checking. If the batch fails, we convert to verified attestations with individual signature checking enabled. Ultimately, we optimistically try to do a batch verification of attestation signatures and fall-back to individual verification if it fails. This opens an attach vector for "poisoning" the attestations and causing us to waste a batch verification. I argue that peer scoring should do a good-enough job of defending against this and the typical-case gains massively outweigh the worst-case losses. ## Additional Info Before this PR, attestation verification took the attestations by value (instead of by reference). It turns out that this was unnecessary and, in my opinion, resulted in some undesirable ergonomics (e.g., we had to pass the attestation back in the `Err` variant to avoid clones). In this PR I've modified attestation verification so that it now takes a reference. I refactored the `beacon_chain/tests/attestation_verification.rs` tests so they use a builder-esque "tester" struct instead of a weird macro. It made it easier for me to test individual/batch with the same set of tests and I think it was a nice tidy-up. Notably, I did this last to try and make sure my new refactors to *actual* production code would pass under the existing test suite.
2021-09-22 08:49:41 +00:00
Ok(())
}
/// Accepts a `VerifiedSyncContribution` and attempts to apply it to `self.op_pool`.
///
/// The op pool is used by local block producers to pack blocks with operations.
pub fn add_contribution_to_block_inclusion_pool(
&self,
contribution: VerifiedSyncContribution<T>,
) -> Result<(), SyncCommitteeError> {
let _timer = metrics::start_timer(&metrics::SYNC_CONTRIBUTION_PROCESSING_APPLY_TO_OP_POOL);
// If there's no eth1 chain then it's impossible to produce blocks and therefore
// useless to put things in the op pool.
if self.eth1_chain.is_some() {
self.op_pool
.insert_sync_contribution(contribution.contribution())
.map_err(Error::from)?;
}
Ok(())
}
/// Filter an attestation from the op pool for shuffling compatibility.
///
/// Use the provided `filter_cache` map to memoize results.
pub fn filter_op_pool_attestation(
&self,
filter_cache: &mut HashMap<(Hash256, Epoch), bool>,
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
att: &AttestationRef<T::EthSpec>,
state: &BeaconState<T::EthSpec>,
) -> bool {
*filter_cache
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
.entry((att.data.beacon_block_root, att.checkpoint.target_epoch))
.or_insert_with(|| {
self.shuffling_is_compatible(
&att.data.beacon_block_root,
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
att.checkpoint.target_epoch,
state,
)
})
}
/// Check that the shuffling at `block_root` is equal to one of the shufflings of `state`.
///
/// The `target_epoch` argument determines which shuffling to check compatibility with, it
/// should be equal to the current or previous epoch of `state`, or else `false` will be
/// returned.
///
/// The compatibility check is designed to be fast: we check that the block that
/// determined the RANDAO mix for the `target_epoch` matches the ancestor of the block
/// identified by `block_root` (at that slot).
pub fn shuffling_is_compatible(
&self,
block_root: &Hash256,
target_epoch: Epoch,
state: &BeaconState<T::EthSpec>,
) -> bool {
Fix attestation shuffling filter (#3629) ## Issue Addressed Fix a bug in block production that results in blocks with 0 attestations during the first slot of an epoch. The bug is marked by debug logs of the form: > DEBG Discarding attestation because of missing ancestor, block_root: 0x3cc00d9c9e0883b2d0db8606278f2b8423d4902f9a1ee619258b5b60590e64f8, pivot_slot: 4042591 It occurs when trying to look up the shuffling decision root for an attestation from a slot which is prior to fork choice's finalized block. This happens frequently when proposing in the first slot of the epoch where we have: - `current_epoch == n` - `attestation.data.target.epoch == n - 1` - attestation shuffling epoch `== n - 3` (decision block being the last block of `n - 3`) - `state.finalized_checkpoint.epoch == n - 2` (first block of `n - 2` is finalized) Hence the shuffling decision slot is out of range of the fork choice backwards iterator _by a single slot_. Unfortunately this bug was hidden when we weren't pruning fork choice, and then reintroduced in v2.5.1 when we fixed the pruning (https://github.com/sigp/lighthouse/releases/tag/v2.5.1). There's no way to turn that off or disable the filtering in our current release, so we need a new release to fix this issue. Fortunately, it also does not occur on every epoch boundary because of the gradual pruning of fork choice every 256 blocks (~8 epochs): https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array_fork_choice.rs#L16 https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array.rs#L713-L716 So the probability of proposing a 0-attestation block given a proposal assignment is approximately `1/32 * 1/8 = 0.39%`. ## Proposed Changes - Load the block's shuffling ID from fork choice and verify it against the expected shuffling ID of the head state. This code was initially written before we had settled on a representation of shuffling IDs, so I think it's a nice simplification to make use of them here rather than more ad-hoc logic that fundamentally does the same thing. ## Additional Info Thanks to @moshe-blox for noticing this issue and bringing it to our attention.
2022-10-18 04:02:06 +00:00
self.shuffling_is_compatible_result(block_root, target_epoch, state)
.unwrap_or_else(|e| {
debug!(
self.log,
"Skipping attestation with incompatible shuffling";
"block_root" => ?block_root,
"target_epoch" => target_epoch,
"reason" => ?e,
);
Fix attestation shuffling filter (#3629) ## Issue Addressed Fix a bug in block production that results in blocks with 0 attestations during the first slot of an epoch. The bug is marked by debug logs of the form: > DEBG Discarding attestation because of missing ancestor, block_root: 0x3cc00d9c9e0883b2d0db8606278f2b8423d4902f9a1ee619258b5b60590e64f8, pivot_slot: 4042591 It occurs when trying to look up the shuffling decision root for an attestation from a slot which is prior to fork choice's finalized block. This happens frequently when proposing in the first slot of the epoch where we have: - `current_epoch == n` - `attestation.data.target.epoch == n - 1` - attestation shuffling epoch `== n - 3` (decision block being the last block of `n - 3`) - `state.finalized_checkpoint.epoch == n - 2` (first block of `n - 2` is finalized) Hence the shuffling decision slot is out of range of the fork choice backwards iterator _by a single slot_. Unfortunately this bug was hidden when we weren't pruning fork choice, and then reintroduced in v2.5.1 when we fixed the pruning (https://github.com/sigp/lighthouse/releases/tag/v2.5.1). There's no way to turn that off or disable the filtering in our current release, so we need a new release to fix this issue. Fortunately, it also does not occur on every epoch boundary because of the gradual pruning of fork choice every 256 blocks (~8 epochs): https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array_fork_choice.rs#L16 https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array.rs#L713-L716 So the probability of proposing a 0-attestation block given a proposal assignment is approximately `1/32 * 1/8 = 0.39%`. ## Proposed Changes - Load the block's shuffling ID from fork choice and verify it against the expected shuffling ID of the head state. This code was initially written before we had settled on a representation of shuffling IDs, so I think it's a nice simplification to make use of them here rather than more ad-hoc logic that fundamentally does the same thing. ## Additional Info Thanks to @moshe-blox for noticing this issue and bringing it to our attention.
2022-10-18 04:02:06 +00:00
false
})
}
Fix attestation shuffling filter (#3629) ## Issue Addressed Fix a bug in block production that results in blocks with 0 attestations during the first slot of an epoch. The bug is marked by debug logs of the form: > DEBG Discarding attestation because of missing ancestor, block_root: 0x3cc00d9c9e0883b2d0db8606278f2b8423d4902f9a1ee619258b5b60590e64f8, pivot_slot: 4042591 It occurs when trying to look up the shuffling decision root for an attestation from a slot which is prior to fork choice's finalized block. This happens frequently when proposing in the first slot of the epoch where we have: - `current_epoch == n` - `attestation.data.target.epoch == n - 1` - attestation shuffling epoch `== n - 3` (decision block being the last block of `n - 3`) - `state.finalized_checkpoint.epoch == n - 2` (first block of `n - 2` is finalized) Hence the shuffling decision slot is out of range of the fork choice backwards iterator _by a single slot_. Unfortunately this bug was hidden when we weren't pruning fork choice, and then reintroduced in v2.5.1 when we fixed the pruning (https://github.com/sigp/lighthouse/releases/tag/v2.5.1). There's no way to turn that off or disable the filtering in our current release, so we need a new release to fix this issue. Fortunately, it also does not occur on every epoch boundary because of the gradual pruning of fork choice every 256 blocks (~8 epochs): https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array_fork_choice.rs#L16 https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array.rs#L713-L716 So the probability of proposing a 0-attestation block given a proposal assignment is approximately `1/32 * 1/8 = 0.39%`. ## Proposed Changes - Load the block's shuffling ID from fork choice and verify it against the expected shuffling ID of the head state. This code was initially written before we had settled on a representation of shuffling IDs, so I think it's a nice simplification to make use of them here rather than more ad-hoc logic that fundamentally does the same thing. ## Additional Info Thanks to @moshe-blox for noticing this issue and bringing it to our attention.
2022-10-18 04:02:06 +00:00
fn shuffling_is_compatible_result(
&self,
block_root: &Hash256,
target_epoch: Epoch,
state: &BeaconState<T::EthSpec>,
) -> Result<bool, Error> {
// Compute the shuffling ID for the head state in the `target_epoch`.
let relative_epoch = RelativeEpoch::from_epoch(state.current_epoch(), target_epoch)
.map_err(|e| Error::BeaconStateError(e.into()))?;
let head_shuffling_id =
AttestationShufflingId::new(self.genesis_block_root, state, relative_epoch)?;
// Load the block's shuffling ID from fork choice. We use the variant of `get_block` that
// checks descent from the finalized block, so there's one case where we'll spuriously
// return `false`: where an attestation for the previous epoch nominates the pivot block
// which is the parent block of the finalized block. Such attestations are not useful, so
// this doesn't matter.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let fork_choice_lock = self.canonical_head.fork_choice_read_lock();
Fix attestation shuffling filter (#3629) ## Issue Addressed Fix a bug in block production that results in blocks with 0 attestations during the first slot of an epoch. The bug is marked by debug logs of the form: > DEBG Discarding attestation because of missing ancestor, block_root: 0x3cc00d9c9e0883b2d0db8606278f2b8423d4902f9a1ee619258b5b60590e64f8, pivot_slot: 4042591 It occurs when trying to look up the shuffling decision root for an attestation from a slot which is prior to fork choice's finalized block. This happens frequently when proposing in the first slot of the epoch where we have: - `current_epoch == n` - `attestation.data.target.epoch == n - 1` - attestation shuffling epoch `== n - 3` (decision block being the last block of `n - 3`) - `state.finalized_checkpoint.epoch == n - 2` (first block of `n - 2` is finalized) Hence the shuffling decision slot is out of range of the fork choice backwards iterator _by a single slot_. Unfortunately this bug was hidden when we weren't pruning fork choice, and then reintroduced in v2.5.1 when we fixed the pruning (https://github.com/sigp/lighthouse/releases/tag/v2.5.1). There's no way to turn that off or disable the filtering in our current release, so we need a new release to fix this issue. Fortunately, it also does not occur on every epoch boundary because of the gradual pruning of fork choice every 256 blocks (~8 epochs): https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array_fork_choice.rs#L16 https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array.rs#L713-L716 So the probability of proposing a 0-attestation block given a proposal assignment is approximately `1/32 * 1/8 = 0.39%`. ## Proposed Changes - Load the block's shuffling ID from fork choice and verify it against the expected shuffling ID of the head state. This code was initially written before we had settled on a representation of shuffling IDs, so I think it's a nice simplification to make use of them here rather than more ad-hoc logic that fundamentally does the same thing. ## Additional Info Thanks to @moshe-blox for noticing this issue and bringing it to our attention.
2022-10-18 04:02:06 +00:00
let block = fork_choice_lock
.get_block(block_root)
.ok_or(Error::AttestationHeadNotInForkChoice(*block_root))?;
drop(fork_choice_lock);
Fix attestation shuffling filter (#3629) ## Issue Addressed Fix a bug in block production that results in blocks with 0 attestations during the first slot of an epoch. The bug is marked by debug logs of the form: > DEBG Discarding attestation because of missing ancestor, block_root: 0x3cc00d9c9e0883b2d0db8606278f2b8423d4902f9a1ee619258b5b60590e64f8, pivot_slot: 4042591 It occurs when trying to look up the shuffling decision root for an attestation from a slot which is prior to fork choice's finalized block. This happens frequently when proposing in the first slot of the epoch where we have: - `current_epoch == n` - `attestation.data.target.epoch == n - 1` - attestation shuffling epoch `== n - 3` (decision block being the last block of `n - 3`) - `state.finalized_checkpoint.epoch == n - 2` (first block of `n - 2` is finalized) Hence the shuffling decision slot is out of range of the fork choice backwards iterator _by a single slot_. Unfortunately this bug was hidden when we weren't pruning fork choice, and then reintroduced in v2.5.1 when we fixed the pruning (https://github.com/sigp/lighthouse/releases/tag/v2.5.1). There's no way to turn that off or disable the filtering in our current release, so we need a new release to fix this issue. Fortunately, it also does not occur on every epoch boundary because of the gradual pruning of fork choice every 256 blocks (~8 epochs): https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array_fork_choice.rs#L16 https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array.rs#L713-L716 So the probability of proposing a 0-attestation block given a proposal assignment is approximately `1/32 * 1/8 = 0.39%`. ## Proposed Changes - Load the block's shuffling ID from fork choice and verify it against the expected shuffling ID of the head state. This code was initially written before we had settled on a representation of shuffling IDs, so I think it's a nice simplification to make use of them here rather than more ad-hoc logic that fundamentally does the same thing. ## Additional Info Thanks to @moshe-blox for noticing this issue and bringing it to our attention.
2022-10-18 04:02:06 +00:00
let block_shuffling_id = if target_epoch == block.current_epoch_shuffling_id.shuffling_epoch
{
block.current_epoch_shuffling_id
} else if target_epoch == block.next_epoch_shuffling_id.shuffling_epoch {
block.next_epoch_shuffling_id
} else if target_epoch > block.next_epoch_shuffling_id.shuffling_epoch {
AttestationShufflingId {
shuffling_epoch: target_epoch,
shuffling_decision_block: *block_root,
}
Fix attestation shuffling filter (#3629) ## Issue Addressed Fix a bug in block production that results in blocks with 0 attestations during the first slot of an epoch. The bug is marked by debug logs of the form: > DEBG Discarding attestation because of missing ancestor, block_root: 0x3cc00d9c9e0883b2d0db8606278f2b8423d4902f9a1ee619258b5b60590e64f8, pivot_slot: 4042591 It occurs when trying to look up the shuffling decision root for an attestation from a slot which is prior to fork choice's finalized block. This happens frequently when proposing in the first slot of the epoch where we have: - `current_epoch == n` - `attestation.data.target.epoch == n - 1` - attestation shuffling epoch `== n - 3` (decision block being the last block of `n - 3`) - `state.finalized_checkpoint.epoch == n - 2` (first block of `n - 2` is finalized) Hence the shuffling decision slot is out of range of the fork choice backwards iterator _by a single slot_. Unfortunately this bug was hidden when we weren't pruning fork choice, and then reintroduced in v2.5.1 when we fixed the pruning (https://github.com/sigp/lighthouse/releases/tag/v2.5.1). There's no way to turn that off or disable the filtering in our current release, so we need a new release to fix this issue. Fortunately, it also does not occur on every epoch boundary because of the gradual pruning of fork choice every 256 blocks (~8 epochs): https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array_fork_choice.rs#L16 https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array.rs#L713-L716 So the probability of proposing a 0-attestation block given a proposal assignment is approximately `1/32 * 1/8 = 0.39%`. ## Proposed Changes - Load the block's shuffling ID from fork choice and verify it against the expected shuffling ID of the head state. This code was initially written before we had settled on a representation of shuffling IDs, so I think it's a nice simplification to make use of them here rather than more ad-hoc logic that fundamentally does the same thing. ## Additional Info Thanks to @moshe-blox for noticing this issue and bringing it to our attention.
2022-10-18 04:02:06 +00:00
} else {
debug!(
self.log,
"Skipping attestation with incompatible shuffling";
"block_root" => ?block_root,
"target_epoch" => target_epoch,
"reason" => "target epoch less than block epoch"
);
return Ok(false);
};
if head_shuffling_id == block_shuffling_id {
Ok(true)
} else {
debug!(
self.log,
"Skipping attestation with incompatible shuffling";
"block_root" => ?block_root,
"target_epoch" => target_epoch,
"head_shuffling_id" => ?head_shuffling_id,
"block_shuffling_id" => ?block_shuffling_id,
);
Ok(false)
}
}
/// Verify a voluntary exit before allowing it to propagate on the gossip network.
pub fn verify_voluntary_exit_for_gossip(
Update to Spec v0.10 (#817) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Disable/delete two outdated tests * Bump eth1 default vote warn to error * Delete outdated eth1 test Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-02-10 23:19:36 +00:00
&self,
exit: SignedVoluntaryExit,
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
) -> Result<ObservationOutcome<SignedVoluntaryExit, T::EthSpec>, Error> {
// NOTE: this could be more efficient if it avoided cloning the head state
let wall_clock_state = self.wall_clock_state()?;
Ok(self
.observed_voluntary_exits
.lock()
.verify_and_observe(exit, &wall_clock_state, &self.spec)
.map(|exit| {
// this method is called for both API and gossip exits, so this covers all exit events
if let Some(event_handler) = self.event_handler.as_ref() {
if event_handler.has_exit_subscribers() {
if let ObservationOutcome::New(exit) = exit.clone() {
event_handler.register(EventKind::VoluntaryExit(exit.into_inner()));
}
}
}
exit
})?)
}
/// Accept a pre-verified exit and queue it for inclusion in an appropriate block.
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
pub fn import_voluntary_exit(&self, exit: SigVerifiedOp<SignedVoluntaryExit, T::EthSpec>) {
if self.eth1_chain.is_some() {
self.op_pool.insert_voluntary_exit(exit)
}
}
/// Verify a proposer slashing before allowing it to propagate on the gossip network.
pub fn verify_proposer_slashing_for_gossip(
&self,
proposer_slashing: ProposerSlashing,
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
) -> Result<ObservationOutcome<ProposerSlashing, T::EthSpec>, Error> {
let wall_clock_state = self.wall_clock_state()?;
Ok(self.observed_proposer_slashings.lock().verify_and_observe(
proposer_slashing,
&wall_clock_state,
&self.spec,
)?)
}
/// Accept some proposer slashing and queue it for inclusion in an appropriate block.
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
pub fn import_proposer_slashing(
&self,
proposer_slashing: SigVerifiedOp<ProposerSlashing, T::EthSpec>,
) {
if self.eth1_chain.is_some() {
self.op_pool.insert_proposer_slashing(proposer_slashing)
}
}
/// Verify an attester slashing before allowing it to propagate on the gossip network.
pub fn verify_attester_slashing_for_gossip(
&self,
Update to frozen spec ❄️ (v0.8.1) (#444) * types: first updates for v0.8 * state_processing: epoch processing v0.8.0 * state_processing: block processing v0.8.0 * tree_hash_derive: support generics in SignedRoot * types v0.8: update to use ssz_types * state_processing v0.8: use ssz_types * ssz_types: add bitwise methods and from_elem * types: fix v0.8 FIXMEs * ssz_types: add bitfield shift_up * ssz_types: iterators and DerefMut for VariableList * types,state_processing: use VariableList * ssz_types: fix BitVector Decode impl Fixed a typo in the implementation of ssz::Decode for BitVector, which caused it to be considered variable length! * types: fix test modules for v0.8 update * types: remove slow type-level arithmetic * state_processing: fix tests for v0.8 * op_pool: update for v0.8 * ssz_types: Bitfield difference length-independent Allow computing the difference of two bitfields of different lengths. * Implement compact committee support * epoch_processing: committee & active index roots * state_processing: genesis state builder v0.8 * state_processing: implement v0.8.1 * Further improve tree_hash * Strip examples, tests from cached_tree_hash * Update TreeHash, un-impl CachedTreeHash * Update bitfield TreeHash, un-impl CachedTreeHash * Update FixedLenVec TreeHash, unimpl CachedTreeHash * Update update tree_hash_derive for new TreeHash * Fix TreeHash, un-impl CachedTreeHash for ssz_types * Remove fixed_len_vec, ssz benches SSZ benches relied upon fixed_len_vec -- it is easier to just delete them and rebuild them later (when necessary) * Remove boolean_bitfield crate * Fix fake_crypto BLS compile errors * Update ef_tests for new v.8 type params * Update ef_tests submodule to v0.8.1 tag * Make fixes to support parsing ssz ef_tests * `compact_committee...` to `compact_committees...` * Derive more traits for `CompactCommittee` * Flip bitfield byte-endianness * Fix tree_hash for bitfields * Modify CLI output for ef_tests * Bump ssz crate version * Update ssz_types doc comment * Del cached tree hash tests from ssz_static tests * Tidy SSZ dependencies * Rename ssz_types crate to eth2_ssz_types * validator_client: update for v0.8 * ssz_types: update union/difference for bit order swap * beacon_node: update for v0.8, EthSpec * types: disable cached tree hash, update min spec * state_processing: fix slot bug in committee update * tests: temporarily disable fork choice harness test See #447 * committee cache: prevent out-of-bounds access In the case where we tried to access the committee of a shard that didn't have a committee in the current epoch, we were accessing elements beyond the end of the shuffling vector and panicking! This commit adds a check to make the failure safe and explicit. * fix bug in get_indexed_attestation and simplify There was a bug in our implementation of get_indexed_attestation whereby incorrect "committee indices" were used to index into the custody bitfield. The bug was only observable in the case where some bits of the custody bitfield were set to 1. The implementation has been simplified to remove the bug, and a test added. * state_proc: workaround for compact committees bug https://github.com/ethereum/eth2.0-specs/issues/1315 * v0.8: updates to make the EF tests pass * Remove redundant max operation checks. * Always supply both messages when checking attestation signatures -- allowing verification of an attestation with no signatures. * Swap the order of the fork and domain constant in `get_domain`, to match the spec. * rustfmt * ef_tests: add new epoch processing tests * Integrate v0.8 into master (compiles) * Remove unused crates, fix clippy lints * Replace v0.6.3 tags w/ v0.8.1 * Remove old comment * Ensure lmd ghost tests only run in release * Update readme
2019-07-30 02:44:51 +00:00
attester_slashing: AttesterSlashing<T::EthSpec>,
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
) -> Result<ObservationOutcome<AttesterSlashing<T::EthSpec>, T::EthSpec>, Error> {
let wall_clock_state = self.wall_clock_state()?;
Ok(self.observed_attester_slashings.lock().verify_and_observe(
attester_slashing,
&wall_clock_state,
&self.spec,
)?)
}
/// Accept a verified attester slashing and:
///
/// 1. Apply it to fork choice.
/// 2. Add it to the op pool.
pub fn import_attester_slashing(
&self,
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
attester_slashing: SigVerifiedOp<AttesterSlashing<T::EthSpec>, T::EthSpec>,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
) {
// Add to fork choice.
self.canonical_head
.fork_choice_write_lock()
.on_attester_slashing(attester_slashing.as_inner());
// Add to the op pool (if we have the ability to propose blocks).
if self.eth1_chain.is_some() {
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
self.op_pool.insert_attester_slashing(attester_slashing)
}
}
/// Attempt to obtain sync committee duties from the head.
pub fn sync_committee_duties_from_head(
&self,
epoch: Epoch,
validator_indices: &[u64],
) -> Result<Vec<Option<SyncDuty>>, Error> {
self.with_head(move |head| {
head.beacon_state
.get_sync_committee_duties(epoch, validator_indices, &self.spec)
.map_err(Error::SyncDutiesError)
})
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
/// A convenience method for spawning a blocking task. It maps an `Option` and
/// `tokio::JoinError` into a single `BeaconChainError`.
pub(crate) async fn spawn_blocking_handle<F, R>(
&self,
task: F,
name: &'static str,
) -> Result<R, Error>
where
F: FnOnce() -> R + Send + 'static,
R: Send + 'static,
{
let handle = self
.task_executor
.spawn_blocking_handle(task, name)
.ok_or(Error::RuntimeShutdown)?;
handle.await.map_err(Error::TokioJoin)
}
/// Accepts a `chain_segment` and filters out any uninteresting blocks (e.g., pre-finalization
/// or already-known).
///
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
/// This method is potentially long-running and should not run on the core executor.
pub fn filter_chain_segment(
self: &Arc<Self>,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
chain_segment: Vec<Arc<SignedBeaconBlock<T::EthSpec>>>,
) -> Result<Vec<HashBlockTuple<T::EthSpec>>, ChainSegmentResult<T::EthSpec>> {
// This function will never import any blocks.
let imported_blocks = 0;
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
let mut filtered_chain_segment = Vec::with_capacity(chain_segment.len());
2019-09-03 04:18:45 +00:00
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// Produce a list of the parent root and slot of the child of each block.
Add proto_array fork choice (#804) * Start implementing proto_array * Add progress * Add unfinished progress * Add further progress * Add progress * Add tree filtering * Add half-finished modifications * Add refactored version * Tidy, add incomplete LmdGhost impl * Move impls in LmdGhost trait def * Remove old reduced_tree fork choice * Combine two functions in to `compute_deltas` * Start testing * Add more compute_deltas tests * Add fork choice testing * Add more fork choice testing * Add more fork choice tests * Add more testing to proto-array * Remove old tests * Modify tests * Add more tests * Add more testing * Add comments and fixes * Re-organise crate * Tidy, finish pruning tests * Add ssz encoding, other pub fns * Rename lmd_ghost > proto_array_fork_choice * Integrate proto_array into lighthouse * Add first pass at fixing filter * Clean out old comments * Add more comments * Attempt to fix prune error * Adjust TODO * Fix test compile errors * Add extra justification change check * Update cargo.lock * Fix fork choice test compile errors * Most remove ffg_update_required * Fix bug with epoch of attestation votes * Start adding new test format * Make fork choice tests declarative * Create test def concept * Move test defs into crate * Add binary, re-org crate * Shuffle files * Start adding ffg tests * Add more fork choice tests * Add fork choice JSON dumping * Add more detail to best node error * Ensure fin+just checkpoints from from same block * Rename JustificationManager * Move checkpoint manager into own file * Tidy * Add targetted logging for sneaky sync bug * Fix justified balances bug * Add cache metrics * Add metrics for log levels * Fix bug in checkpoint manager * Fix compile error in fork choice tests * Ignore duplicate blocks in fork choice * Add block to fock choice before db * Rename on_new_block fn * Fix spec inconsistency in `CheckpointManager` * Remove BlockRootTree * Remove old reduced_tree code fragment * Add API endpoint for fork choice * Add more ffg tests * Remove block_root_tree reminents * Ensure effective balances are used * Remove old debugging code, fix API fault * Add check to ensure parent block is in fork choice * Update readme dates * Fix readme * Tidy checkpoint manager * Remove fork choice yaml files from repo * Remove fork choice yaml from repo * General tidy * Address majority of Michael's comments * Tidy bin/lib business * Remove dangling file * Undo changes for rpc/handler from master * Revert "Undo changes for rpc/handler from master" This reverts commit 876edff0e4a501aafbb47113454852826dcc24e8. Co-authored-by: Age Manning <Age@AgeManning.com>
2020-01-29 04:05:00 +00:00
//
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// E.g., `children[0] == (chain_segment[1].parent_root(), chain_segment[1].slot())`
let children = chain_segment
.iter()
.skip(1)
.map(|block| (block.parent_root(), block.slot()))
.collect::<Vec<_>>();
for (i, block) in chain_segment.into_iter().enumerate() {
// Ensure the block is the correct structure for the fork at `block.slot()`.
if let Err(e) = block.fork_name(&self.spec) {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
return Err(ChainSegmentResult::Failed {
imported_blocks,
error: BlockError::InconsistentFork(e),
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
});
}
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
let block_root = get_block_root(&block);
if let Some((child_parent_root, child_slot)) = children.get(i) {
// If this block has a child in this chain segment, ensure that its parent root matches
// the root of this block.
//
// Without this check it would be possible to have a block verified using the
// incorrect shuffling. That would be bad, mmkay.
if block_root != *child_parent_root {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
return Err(ChainSegmentResult::Failed {
imported_blocks,
error: BlockError::NonLinearParentRoots,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
});
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
}
// Ensure that the slots are strictly increasing throughout the chain segment.
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
if *child_slot <= block.slot() {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
return Err(ChainSegmentResult::Failed {
imported_blocks,
error: BlockError::NonLinearSlots,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
});
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
}
}
Fix block processing blowup, upgrade metrics (#500) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Begin metrics refactor * Move beacon_chain to new metrics structure. * Make metrics not panic if already defined * Use global prometheus gather at rest api * Unify common metric fns into a crate * Add heavy metering to block processing * Remove hypen from prometheus metric name * Add more beacon chain metrics * Add beacon chain persistence metric * Prune op pool on finalization * Add extra prom beacon chain metrics * Prefix BeaconChain metrics with "beacon_" * Add more store metrics * Add basic metrics to libp2p * Add metrics to HTTP server * Remove old `http_server` crate * Update metrics names to be more like standard * Fix broken beacon chain metrics, add slot clock metrics * Add lighthouse_metrics gather fn * Remove http args * Fix wrong state given to op pool prune * Make prom metric names more consistent * Add more metrics, tidy existing metrics * Fix store block read metrics * Tidy attestation metrics * Fix minor PR comments * Allow travis failures on beta (see desc) There's a non-backward compatible change in `cargo fmt`. Stable and beta do not agree. * Tidy `lighthouse_metrics` docs * Fix typo
2019-08-19 11:02:34 +00:00
match check_block_relevancy(&block, block_root, self) {
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// If the block is relevant, add it to the filtered chain segment.
Ok(_) => filtered_chain_segment.push((block_root, block)),
// If the block is already known, simply ignore this block.
Err(BlockError::BlockIsAlreadyKnown) => continue,
// If the block is the genesis block, simply ignore this block.
Err(BlockError::GenesisBlock) => continue,
// If the block is is for a finalized slot, simply ignore this block.
//
// The block is either:
//
// 1. In the canonical finalized chain.
// 2. In some non-canonical chain at a slot that has been finalized already.
//
// In the case of (1), there's no need to re-import and later blocks in this
// segement might be useful.
//
// In the case of (2), skipping the block is valid since we should never import it.
// However, we will potentially get a `ParentUnknown` on a later block. The sync
// protocol will need to ensure this is handled gracefully.
Err(BlockError::WouldRevertFinalizedSlot { .. }) => continue,
// The block has a known parent that does not descend from the finalized block.
// There is no need to process this block or any children.
Err(BlockError::NotFinalizedDescendant { block_parent_root }) => {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
return Err(ChainSegmentResult::Failed {
imported_blocks,
error: BlockError::NotFinalizedDescendant { block_parent_root },
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
});
}
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// If there was an error whilst determining if the block was invalid, return that
// error.
Err(BlockError::BeaconChainError(e)) => {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
return Err(ChainSegmentResult::Failed {
imported_blocks,
error: BlockError::BeaconChainError(e),
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
});
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
}
// If the block was decided to be irrelevant for any other reason, don't include
// this block or any of it's children in the filtered chain segment.
_ => break,
}
2019-06-07 06:46:07 +00:00
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
Ok(filtered_chain_segment)
}
/// Attempt to verify and import a chain of blocks to `self`.
///
/// The provided blocks _must_ each reference the previous block via `block.parent_root` (i.e.,
/// be a chain). An error will be returned if this is not the case.
///
/// This operation is not atomic; if one of the blocks in the chain is invalid then some prior
/// blocks might be imported.
///
/// This method is generally much more efficient than importing each block using
/// `Self::process_block`.
pub async fn process_chain_segment(
self: &Arc<Self>,
chain_segment: Vec<Arc<SignedBeaconBlock<T::EthSpec>>>,
count_unrealized: CountUnrealized,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
) -> ChainSegmentResult<T::EthSpec> {
let mut imported_blocks = 0;
// Filter uninteresting blocks from the chain segment in a blocking task.
let chain = self.clone();
let filtered_chain_segment_future = self.spawn_blocking_handle(
move || chain.filter_chain_segment(chain_segment),
"filter_chain_segment",
);
let mut filtered_chain_segment = match filtered_chain_segment_future.await {
Ok(Ok(filtered_segment)) => filtered_segment,
Ok(Err(segment_result)) => return segment_result,
Err(error) => {
return ChainSegmentResult::Failed {
imported_blocks,
error: BlockError::BeaconChainError(error),
}
}
};
while let Some((_root, block)) = filtered_chain_segment.first() {
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// Determine the epoch of the first block in the remaining segment.
let start_epoch = block.slot().epoch(T::EthSpec::slots_per_epoch());
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// The `last_index` indicates the position of the first block in an epoch greater
// than the current epoch: partitioning the blocks into a run of blocks in the same
// epoch and everything else. These same-epoch blocks can all be signature-verified with
// the same `BeaconState`.
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
let last_index = filtered_chain_segment
.iter()
.position(|(_root, block)| {
block.slot().epoch(T::EthSpec::slots_per_epoch()) > start_epoch
})
.unwrap_or(filtered_chain_segment.len());
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
let mut blocks = filtered_chain_segment.split_off(last_index);
std::mem::swap(&mut blocks, &mut filtered_chain_segment);
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let chain = self.clone();
let signature_verification_future = self.spawn_blocking_handle(
move || signature_verify_chain_segment(blocks, &chain),
"signature_verify_chain_segment",
);
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// Verify the signature of the blocks, returning early if the signature is invalid.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let signature_verified_blocks = match signature_verification_future.await {
Ok(Ok(blocks)) => blocks,
Ok(Err(error)) => {
return ChainSegmentResult::Failed {
imported_blocks,
error,
};
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
Err(error) => {
return ChainSegmentResult::Failed {
imported_blocks,
error: BlockError::BeaconChainError(error),
};
}
};
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// Import the blocks into the chain.
for signature_verified_block in signature_verified_blocks {
match self
.process_block(
signature_verified_block.block_root(),
signature_verified_block,
count_unrealized,
)
.await
{
Ok(_) => imported_blocks += 1,
Err(error) => {
return ChainSegmentResult::Failed {
imported_blocks,
error,
};
}
}
}
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
}
Fix block processing blowup, upgrade metrics (#500) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Begin metrics refactor * Move beacon_chain to new metrics structure. * Make metrics not panic if already defined * Use global prometheus gather at rest api * Unify common metric fns into a crate * Add heavy metering to block processing * Remove hypen from prometheus metric name * Add more beacon chain metrics * Add beacon chain persistence metric * Prune op pool on finalization * Add extra prom beacon chain metrics * Prefix BeaconChain metrics with "beacon_" * Add more store metrics * Add basic metrics to libp2p * Add metrics to HTTP server * Remove old `http_server` crate * Update metrics names to be more like standard * Fix broken beacon chain metrics, add slot clock metrics * Add lighthouse_metrics gather fn * Remove http args * Fix wrong state given to op pool prune * Make prom metric names more consistent * Add more metrics, tidy existing metrics * Fix store block read metrics * Tidy attestation metrics * Fix minor PR comments * Allow travis failures on beta (see desc) There's a non-backward compatible change in `cargo fmt`. Stable and beta do not agree. * Tidy `lighthouse_metrics` docs * Fix typo
2019-08-19 11:02:34 +00:00
ChainSegmentResult::Successful { imported_blocks }
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
}
2019-09-09 00:55:15 +00:00
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
/// Returns `Ok(GossipVerifiedBlock)` if the supplied `block` should be forwarded onto the
/// gossip network. The block is not imported into the chain, it is just partially verified.
///
/// The returned `GossipVerifiedBlock` should be provided to `Self::process_block` immediately
/// after it is returned, unless some other circumstance decides it should not be imported at
/// all.
///
/// ## Errors
///
/// Returns an `Err` if the given block was invalid, or an error was encountered during
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub async fn verify_block_for_gossip(
self: &Arc<Self>,
block: Arc<SignedBeaconBlock<T::EthSpec>>,
) -> Result<GossipVerifiedBlock<T>, BlockError<T::EthSpec>> {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let chain = self.clone();
self.task_executor
.clone()
.spawn_blocking_handle(
move || {
let slot = block.slot();
let graffiti_string = block.message().body().graffiti().as_utf8_lossy();
match GossipVerifiedBlock::new(block, &chain) {
Ok(verified) => {
debug!(
chain.log,
"Successfully verified gossip block";
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
"graffiti" => graffiti_string,
"slot" => slot,
"root" => ?verified.block_root(),
);
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
Ok(verified)
}
Err(e) => {
debug!(
chain.log,
"Rejected gossip block";
"error" => e.to_string(),
"graffiti" => graffiti_string,
"slot" => slot,
);
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
Err(e)
}
}
},
"payload_verification_handle",
)
.ok_or(BeaconChainError::RuntimeShutdown)?
.await
.map_err(BeaconChainError::TokioJoin)?
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
}
Fix block processing blowup, upgrade metrics (#500) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Begin metrics refactor * Move beacon_chain to new metrics structure. * Make metrics not panic if already defined * Use global prometheus gather at rest api * Unify common metric fns into a crate * Add heavy metering to block processing * Remove hypen from prometheus metric name * Add more beacon chain metrics * Add beacon chain persistence metric * Prune op pool on finalization * Add extra prom beacon chain metrics * Prefix BeaconChain metrics with "beacon_" * Add more store metrics * Add basic metrics to libp2p * Add metrics to HTTP server * Remove old `http_server` crate * Update metrics names to be more like standard * Fix broken beacon chain metrics, add slot clock metrics * Add lighthouse_metrics gather fn * Remove http args * Fix wrong state given to op pool prune * Make prom metric names more consistent * Add more metrics, tidy existing metrics * Fix store block read metrics * Tidy attestation metrics * Fix minor PR comments * Allow travis failures on beta (see desc) There's a non-backward compatible change in `cargo fmt`. Stable and beta do not agree. * Tidy `lighthouse_metrics` docs * Fix typo
2019-08-19 11:02:34 +00:00
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
/// Returns `Ok(block_root)` if the given `unverified_block` was successfully verified and
/// imported into the chain.
///
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
/// Items that implement `IntoExecutionPendingBlock` include:
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
///
/// - `SignedBeaconBlock`
/// - `GossipVerifiedBlock`
///
/// ## Errors
///
/// Returns an `Err` if the given block was invalid, or an error was encountered during
/// verification.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub async fn process_block<B: IntoExecutionPendingBlock<T>>(
self: &Arc<Self>,
block_root: Hash256,
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
unverified_block: B,
count_unrealized: CountUnrealized,
) -> Result<Hash256, BlockError<T::EthSpec>> {
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// Start the Prometheus timer.
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
let _full_timer = metrics::start_timer(&metrics::BLOCK_PROCESSING_TIMES);
Bootstrap (#501) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Begin metrics refactor * Move beacon_chain to new metrics structure. * Make metrics not panic if already defined * Use global prometheus gather at rest api * Unify common metric fns into a crate * Add heavy metering to block processing * Remove hypen from prometheus metric name * Add more beacon chain metrics * Add beacon chain persistence metric * Prune op pool on finalization * Add extra prom beacon chain metrics * Prefix BeaconChain metrics with "beacon_" * Add more store metrics * Add basic metrics to libp2p * Add metrics to HTTP server * Remove old `http_server` crate * Update metrics names to be more like standard * Fix broken beacon chain metrics, add slot clock metrics * Add lighthouse_metrics gather fn * Remove http args * Fix wrong state given to op pool prune * Make prom metric names more consistent * Add more metrics, tidy existing metrics * Fix store block read metrics * Tidy attestation metrics * Fix minor PR comments * Fix minor PR comments * Remove duplicated attestation finalization check * Remove awkward `let` statement * Add first attempts at HTTP bootstrap * Add beacon_block methods to rest api * Fix serde for block.body.grafitti * Allow travis failures on beta (see desc) There's a non-backward compatible change in `cargo fmt`. Stable and beta do not agree. * Add network routes to API * Fix rustc warnings * Add best_slot method * Add --bootstrap arg to beacon node * Get bootstrapper working for ENR address * Store intermediate states during block processing * Allow bootstrapper to scrape libp2p address * Update bootstrapper libp2p address finding * Add comments * Tidy API to be more consistent with recent decisions * Address some review comments * Make BeaconChainTypes Send + Sync + 'static * Add `/network/listen_port` API endpoint * Abandon starting the node if libp2p doesn't start * Update bootstrapper for API changes * Remove unnecessary trait bounds
2019-08-23 05:53:53 +00:00
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// Increment the Prometheus counter for block processing requests.
metrics::inc_counter(&metrics::BLOCK_PROCESSING_REQUESTS);
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// Clone the block so we can provide it to the event handler.
let block = unverified_block.block().clone();
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// A small closure to group the verification and import errors.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let chain = self.clone();
let import_block = async move {
let execution_pending =
unverified_block.into_execution_pending_block(block_root, &chain)?;
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
chain
.import_execution_pending_block(execution_pending, count_unrealized)
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
.await
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
};
Fix block processing blowup, upgrade metrics (#500) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Begin metrics refactor * Move beacon_chain to new metrics structure. * Make metrics not panic if already defined * Use global prometheus gather at rest api * Unify common metric fns into a crate * Add heavy metering to block processing * Remove hypen from prometheus metric name * Add more beacon chain metrics * Add beacon chain persistence metric * Prune op pool on finalization * Add extra prom beacon chain metrics * Prefix BeaconChain metrics with "beacon_" * Add more store metrics * Add basic metrics to libp2p * Add metrics to HTTP server * Remove old `http_server` crate * Update metrics names to be more like standard * Fix broken beacon chain metrics, add slot clock metrics * Add lighthouse_metrics gather fn * Remove http args * Fix wrong state given to op pool prune * Make prom metric names more consistent * Add more metrics, tidy existing metrics * Fix store block read metrics * Tidy attestation metrics * Fix minor PR comments * Allow travis failures on beta (see desc) There's a non-backward compatible change in `cargo fmt`. Stable and beta do not agree. * Tidy `lighthouse_metrics` docs * Fix typo
2019-08-19 11:02:34 +00:00
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// Verify and import the block.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
match import_block.await {
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// The block was successfully verified and imported. Yay.
Ok(block_root) => {
trace!(
self.log,
"Beacon block imported";
"block_root" => ?block_root,
"block_slot" => %block.slot(),
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
);
Fix block processing blowup, upgrade metrics (#500) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Begin metrics refactor * Move beacon_chain to new metrics structure. * Make metrics not panic if already defined * Use global prometheus gather at rest api * Unify common metric fns into a crate * Add heavy metering to block processing * Remove hypen from prometheus metric name * Add more beacon chain metrics * Add beacon chain persistence metric * Prune op pool on finalization * Add extra prom beacon chain metrics * Prefix BeaconChain metrics with "beacon_" * Add more store metrics * Add basic metrics to libp2p * Add metrics to HTTP server * Remove old `http_server` crate * Update metrics names to be more like standard * Fix broken beacon chain metrics, add slot clock metrics * Add lighthouse_metrics gather fn * Remove http args * Fix wrong state given to op pool prune * Make prom metric names more consistent * Add more metrics, tidy existing metrics * Fix store block read metrics * Tidy attestation metrics * Fix minor PR comments * Allow travis failures on beta (see desc) There's a non-backward compatible change in `cargo fmt`. Stable and beta do not agree. * Tidy `lighthouse_metrics` docs * Fix typo
2019-08-19 11:02:34 +00:00
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// Increment the Prometheus counter for block processing successes.
metrics::inc_counter(&metrics::BLOCK_PROCESSING_SUCCESSES);
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
Ok(block_root)
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
Err(e @ BlockError::BeaconChainError(BeaconChainError::TokioJoin(_))) => {
debug!(
self.log,
"Beacon block processing cancelled";
"error" => ?e,
);
Err(e)
}
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// There was an error whilst attempting to verify and import the block. The block might
// be partially verified or partially imported.
Err(BlockError::BeaconChainError(e)) => {
crit!(
self.log,
"Beacon block processing error";
"error" => ?e,
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
);
Err(BlockError::BeaconChainError(e))
}
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
// The block failed verification.
Err(other) => {
trace!(
self.log,
"Beacon block rejected";
"reason" => other.to_string(),
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
);
Err(other)
}
}
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
}
2019-09-09 00:55:15 +00:00
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
/// Accepts a fully-verified block and imports it into the chain without performing any
/// additional verification.
///
/// An error is returned if the block was unable to be imported. It may be partially imported
/// (i.e., this function is not atomic).
async fn import_execution_pending_block(
self: Arc<Self>,
execution_pending_block: ExecutionPendingBlock<T>,
count_unrealized: CountUnrealized,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
) -> Result<Hash256, BlockError<T::EthSpec>> {
let ExecutionPendingBlock {
block,
block_root,
state,
Deposit Cache Finalization & Fast WS Sync (#2915) ## Summary The deposit cache now has the ability to finalize deposits. This will cause it to drop unneeded deposit logs and hashes in the deposit Merkle tree that are no longer required to construct deposit proofs. The cache is finalized whenever the latest finalized checkpoint has a new `Eth1Data` with all deposits imported. This has three benefits: 1. Improves the speed of constructing Merkle proofs for deposits as we can just replay deposits since the last finalized checkpoint instead of all historical deposits when re-constructing the Merkle tree. 2. Significantly faster weak subjectivity sync as the deposit cache can be transferred to the newly syncing node in compressed form. The Merkle tree that stores `N` finalized deposits requires a maximum of `log2(N)` hashes. The newly syncing node then only needs to download deposits since the last finalized checkpoint to have a full tree. 3. Future proofing in preparation for [EIP-4444](https://eips.ethereum.org/EIPS/eip-4444) as execution nodes will no longer be required to store logs permanently so we won't always have all historical logs available to us. ## More Details Image to illustrate how the deposit contract merkle tree evolves and finalizes along with the resulting `DepositTreeSnapshot` ![image](https://user-images.githubusercontent.com/37123614/151465302-5fc56284-8a69-4998-b20e-45db3934ac70.png) ## Other Considerations I've changed the structure of the `SszDepositCache` so once you load & save your database from this version of lighthouse, you will no longer be able to load it from older versions. Co-authored-by: ethDreamer <37123614+ethDreamer@users.noreply.github.com>
2022-10-30 04:04:24 +00:00
parent_block,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
confirmed_state_roots,
payload_verification_handle,
Deposit Cache Finalization & Fast WS Sync (#2915) ## Summary The deposit cache now has the ability to finalize deposits. This will cause it to drop unneeded deposit logs and hashes in the deposit Merkle tree that are no longer required to construct deposit proofs. The cache is finalized whenever the latest finalized checkpoint has a new `Eth1Data` with all deposits imported. This has three benefits: 1. Improves the speed of constructing Merkle proofs for deposits as we can just replay deposits since the last finalized checkpoint instead of all historical deposits when re-constructing the Merkle tree. 2. Significantly faster weak subjectivity sync as the deposit cache can be transferred to the newly syncing node in compressed form. The Merkle tree that stores `N` finalized deposits requires a maximum of `log2(N)` hashes. The newly syncing node then only needs to download deposits since the last finalized checkpoint to have a full tree. 3. Future proofing in preparation for [EIP-4444](https://eips.ethereum.org/EIPS/eip-4444) as execution nodes will no longer be required to store logs permanently so we won't always have all historical logs available to us. ## More Details Image to illustrate how the deposit contract merkle tree evolves and finalizes along with the resulting `DepositTreeSnapshot` ![image](https://user-images.githubusercontent.com/37123614/151465302-5fc56284-8a69-4998-b20e-45db3934ac70.png) ## Other Considerations I've changed the structure of the `SszDepositCache` so once you load & save your database from this version of lighthouse, you will no longer be able to load it from older versions. Co-authored-by: ethDreamer <37123614+ethDreamer@users.noreply.github.com>
2022-10-30 04:04:24 +00:00
parent_eth1_finalization_data,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
} = execution_pending_block;
let PayloadVerificationOutcome {
payload_verification_status,
is_valid_merge_transition_block,
} = payload_verification_handle
.await
.map_err(BeaconChainError::TokioJoin)?
.ok_or(BeaconChainError::RuntimeShutdown)??;
// Log the PoS pandas if a merge transition just occurred.
if is_valid_merge_transition_block {
info!(self.log, "{}", POS_PANDA_BANNER);
info!(
self.log,
"Proof of Stake Activated";
"slot" => block.slot()
);
info!(
self.log, "";
"Terminal POW Block Hash" => ?block
.message()
.execution_payload()?
.parent_hash()
.into_root()
);
info!(
self.log, "";
"Merge Transition Block Root" => ?block.message().tree_hash_root()
);
info!(
self.log, "";
"Merge Transition Execution Hash" => ?block
.message()
.execution_payload()?
.block_hash()
.into_root()
);
}
let chain = self.clone();
let block_hash = self
.spawn_blocking_handle(
move || {
chain.import_block(
block,
block_root,
state,
confirmed_state_roots,
payload_verification_status,
count_unrealized,
Deposit Cache Finalization & Fast WS Sync (#2915) ## Summary The deposit cache now has the ability to finalize deposits. This will cause it to drop unneeded deposit logs and hashes in the deposit Merkle tree that are no longer required to construct deposit proofs. The cache is finalized whenever the latest finalized checkpoint has a new `Eth1Data` with all deposits imported. This has three benefits: 1. Improves the speed of constructing Merkle proofs for deposits as we can just replay deposits since the last finalized checkpoint instead of all historical deposits when re-constructing the Merkle tree. 2. Significantly faster weak subjectivity sync as the deposit cache can be transferred to the newly syncing node in compressed form. The Merkle tree that stores `N` finalized deposits requires a maximum of `log2(N)` hashes. The newly syncing node then only needs to download deposits since the last finalized checkpoint to have a full tree. 3. Future proofing in preparation for [EIP-4444](https://eips.ethereum.org/EIPS/eip-4444) as execution nodes will no longer be required to store logs permanently so we won't always have all historical logs available to us. ## More Details Image to illustrate how the deposit contract merkle tree evolves and finalizes along with the resulting `DepositTreeSnapshot` ![image](https://user-images.githubusercontent.com/37123614/151465302-5fc56284-8a69-4998-b20e-45db3934ac70.png) ## Other Considerations I've changed the structure of the `SszDepositCache` so once you load & save your database from this version of lighthouse, you will no longer be able to load it from older versions. Co-authored-by: ethDreamer <37123614+ethDreamer@users.noreply.github.com>
2022-10-30 04:04:24 +00:00
parent_block,
parent_eth1_finalization_data,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
)
},
"payload_verification_handle",
)
.await??;
Ok(block_hash)
}
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
/// Accepts a fully-verified block and imports it into the chain without performing any
/// additional verification.
///
/// An error is returned if the block was unable to be imported. It may be partially imported
/// (i.e., this function is not atomic).
Deposit Cache Finalization & Fast WS Sync (#2915) ## Summary The deposit cache now has the ability to finalize deposits. This will cause it to drop unneeded deposit logs and hashes in the deposit Merkle tree that are no longer required to construct deposit proofs. The cache is finalized whenever the latest finalized checkpoint has a new `Eth1Data` with all deposits imported. This has three benefits: 1. Improves the speed of constructing Merkle proofs for deposits as we can just replay deposits since the last finalized checkpoint instead of all historical deposits when re-constructing the Merkle tree. 2. Significantly faster weak subjectivity sync as the deposit cache can be transferred to the newly syncing node in compressed form. The Merkle tree that stores `N` finalized deposits requires a maximum of `log2(N)` hashes. The newly syncing node then only needs to download deposits since the last finalized checkpoint to have a full tree. 3. Future proofing in preparation for [EIP-4444](https://eips.ethereum.org/EIPS/eip-4444) as execution nodes will no longer be required to store logs permanently so we won't always have all historical logs available to us. ## More Details Image to illustrate how the deposit contract merkle tree evolves and finalizes along with the resulting `DepositTreeSnapshot` ![image](https://user-images.githubusercontent.com/37123614/151465302-5fc56284-8a69-4998-b20e-45db3934ac70.png) ## Other Considerations I've changed the structure of the `SszDepositCache` so once you load & save your database from this version of lighthouse, you will no longer be able to load it from older versions. Co-authored-by: ethDreamer <37123614+ethDreamer@users.noreply.github.com>
2022-10-30 04:04:24 +00:00
#[allow(clippy::too_many_arguments)]
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
fn import_block(
&self,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
signed_block: Arc<SignedBeaconBlock<T::EthSpec>>,
block_root: Hash256,
mut state: BeaconState<T::EthSpec>,
confirmed_state_roots: Vec<Hash256>,
payload_verification_status: PayloadVerificationStatus,
count_unrealized: CountUnrealized,
Deposit Cache Finalization & Fast WS Sync (#2915) ## Summary The deposit cache now has the ability to finalize deposits. This will cause it to drop unneeded deposit logs and hashes in the deposit Merkle tree that are no longer required to construct deposit proofs. The cache is finalized whenever the latest finalized checkpoint has a new `Eth1Data` with all deposits imported. This has three benefits: 1. Improves the speed of constructing Merkle proofs for deposits as we can just replay deposits since the last finalized checkpoint instead of all historical deposits when re-constructing the Merkle tree. 2. Significantly faster weak subjectivity sync as the deposit cache can be transferred to the newly syncing node in compressed form. The Merkle tree that stores `N` finalized deposits requires a maximum of `log2(N)` hashes. The newly syncing node then only needs to download deposits since the last finalized checkpoint to have a full tree. 3. Future proofing in preparation for [EIP-4444](https://eips.ethereum.org/EIPS/eip-4444) as execution nodes will no longer be required to store logs permanently so we won't always have all historical logs available to us. ## More Details Image to illustrate how the deposit contract merkle tree evolves and finalizes along with the resulting `DepositTreeSnapshot` ![image](https://user-images.githubusercontent.com/37123614/151465302-5fc56284-8a69-4998-b20e-45db3934ac70.png) ## Other Considerations I've changed the structure of the `SszDepositCache` so once you load & save your database from this version of lighthouse, you will no longer be able to load it from older versions. Co-authored-by: ethDreamer <37123614+ethDreamer@users.noreply.github.com>
2022-10-30 04:04:24 +00:00
parent_block: SignedBlindedBeaconBlock<T::EthSpec>,
parent_eth1_finalization_data: Eth1FinalizationData,
) -> Result<Hash256, BlockError<T::EthSpec>> {
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
let current_slot = self.slot()?;
let current_epoch = current_slot.epoch(T::EthSpec::slots_per_epoch());
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
let attestation_observation_timer =
metrics::start_timer(&metrics::BLOCK_PROCESSING_ATTESTATION_OBSERVATION);
// Iterate through the attestations in the block and register them as an "observed
// attestation". This will stop us from propagating them on the gossip network.
for a in signed_block.message().body().attestations() {
match self.observed_attestations.write().observe_item(a, None) {
Add attestation gossip pre-verification (#983) * Add PH & MS slot clock changes * Account for genesis time * Add progress on duties refactor * Add simple is_aggregator bool to val subscription * Start work on attestation_verification.rs * Add progress on ObservedAttestations * Progress with ObservedAttestations * Fix tests * Add observed attestations to the beacon chain * Add attestation observation to processing code * Add progress on attestation verification * Add first draft of ObservedAttesters * Add more tests * Add observed attesters to beacon chain * Add observers to attestation processing * Add more attestation verification * Create ObservedAggregators map * Remove commented-out code * Add observed aggregators into chain * Add progress * Finish adding features to attestation verification * Ensure beacon chain compiles * Link attn verification into chain * Integrate new attn verification in chain * Remove old attestation processing code * Start trying to fix beacon_chain tests * Split adding into pools into two functions * Add aggregation to harness * Get test harness working again * Adjust the number of aggregators for test harness * Fix edge-case in harness * Integrate new attn processing in network * Fix compile bug in validator_client * Update validator API endpoints * Fix aggreagation in test harness * Fix enum thing * Fix attestation observation bug: * Patch failing API tests * Start adding comments to attestation verification * Remove unused attestation field * Unify "is block known" logic * Update comments * Supress fork choice errors for network processing * Add todos * Tidy * Add gossip attn tests * Disallow test harness to produce old attns * Comment out in-progress tests * Partially address pruning tests * Fix failing store test * Add aggregate tests * Add comments about which spec conditions we check * Dont re-aggregate * Split apart test harness attn production * Fix compile error in network * Make progress on commented-out test * Fix skipping attestation test * Add fork choice verification tests * Tidy attn tests, remove dead code * Remove some accidentally added code * Fix clippy lint * Rename test file * Add block tests, add cheap block proposer check * Rename block testing file * Add observed_block_producers * Tidy * Switch around block signature verification * Finish block testing * Remove gossip from signature tests * First pass of self review * Fix deviation in spec * Update test spec tags * Start moving over to hashset * Finish moving observed attesters to hashmap * Move aggregation pool over to hashmap * Make fc attn borrow again * Fix rest_api compile error * Fix missing comments * Fix monster test * Uncomment increasing slots test * Address remaining comments * Remove unsafe, use cfg test * Remove cfg test flag * Fix dodgy comment * Ignore aggregates that are already known. * Unify aggregator modulo logic * Fix typo in logs * Refactor validator subscription logic * Avoid reproducing selection proof * Skip HTTP call if no subscriptions * Rename DutyAndState -> DutyAndProof * Tidy logs * Print root as dbg * Fix compile errors in tests * Fix compile error in test
2020-05-06 11:42:56 +00:00
// If the observation was successful or if the slot for the attestation was too
// low, continue.
//
// We ignore `SlotTooLow` since this will be very common whilst syncing.
Ok(_) | Err(AttestationObservationError::SlotTooLow { .. }) => {}
Err(e) => return Err(BlockError::BeaconChainError(e.into())),
}
}
metrics::stop_timer(attestation_observation_timer);
// If a slasher is configured, provide the attestations from the block.
if let Some(slasher) = self.slasher.as_ref() {
for attestation in signed_block.message().body().attestations() {
let committee =
state.get_beacon_committee(attestation.data.slot, attestation.data.index)?;
let indexed_attestation = get_indexed_attestation(committee.committee, attestation)
.map_err(|e| BlockError::BeaconChainError(e.into()))?;
slasher.accept_attestation(indexed_attestation);
}
}
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
// If there are new validators in this block, update our pubkey cache.
//
// We perform this _before_ adding the block to fork choice because the pubkey cache is
// used by attestation processing which will only process an attestation if the block is
// known to fork choice. This ordering ensure that the pubkey cache is always up-to-date.
self.validator_pubkey_cache
.try_write_for(VALIDATOR_PUBKEY_CACHE_LOCK_TIMEOUT)
.ok_or(Error::ValidatorPubkeyCacheLockTimeout)?
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
.import_new_pubkeys(&state)?;
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
// For the current and next epoch of this state, ensure we have the shuffling from this
// block in our cache.
for relative_epoch in &[RelativeEpoch::Current, RelativeEpoch::Next] {
Advance state to next slot after importing block (#2174) ## Issue Addressed NA ## Proposed Changes Add an optimization to perform `per_slot_processing` from the *leading-edge* of block processing to the *trailing-edge*. Ultimately, this allows us to import the block at slot `n` faster because we used the tail-end of slot `n - 1` to perform `per_slot_processing`. Additionally, add a "block proposer cache" which allows us to cache the block proposer for some epoch. Since we're now doing trailing-edge `per_slot_processing`, we can prime this cache with the values for the next epoch before those blocks arrive (assuming those blocks don't have some weird forking). There were several ancillary changes required to achieve this: - Remove the `state_root` field of `BeaconSnapshot`, since there's no need to know it on a `pre_state` and in all other cases we can just read it from `block.state_root()`. - This caused some "dust" changes of `snapshot.beacon_state_root` to `snapshot.beacon_state_root()`, where the `BeaconSnapshot::beacon_state_root()` func just reads the state root from the block. - Rename `types::ShuffingId` to `AttestationShufflingId`. I originally did this because I added a `ProposerShufflingId` struct which turned out to be not so useful. I thought this new name was more descriptive so I kept it. - Address https://github.com/ethereum/eth2.0-specs/pull/2196 - Add a debug log when we get a block with an unknown parent. There was previously no logging around this case. - Add a function to `BeaconState` to compute all proposers for an epoch without re-computing the active indices for each slot. ## Additional Info - ~~Blocked on #2173~~ - ~~Blocked on #2179~~ That PR was wrapped into this PR. - There's potentially some places where we could avoid computing the proposer indices in `per_block_processing` but I haven't done this here. These would be an optimization beyond the issue at hand (improving block propagation times) and I think this PR is already doing enough. We can come back for that later. ## TODO - [x] Tidy, improve comments. - [x] ~~Try avoid computing proposer index in `per_block_processing`?~~
2021-02-15 07:17:52 +00:00
let shuffling_id = AttestationShufflingId::new(block_root, &state, *relative_epoch)?;
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
let shuffling_is_cached = self
.shuffling_cache
.try_read_for(ATTESTATION_CACHE_LOCK_TIMEOUT)
.ok_or(Error::AttestationCacheLockTimeout)?
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
.contains(&shuffling_id);
if !shuffling_is_cached {
state.build_committee_cache(*relative_epoch, &self.spec)?;
let committee_cache = state.committee_cache(*relative_epoch)?;
self.shuffling_cache
.try_write_for(ATTESTATION_CACHE_LOCK_TIMEOUT)
.ok_or(Error::AttestationCacheLockTimeout)?
Avoid duplicate committee cache loads (#3574) ## Issue Addressed NA ## Proposed Changes I have observed scenarios on Goerli where Lighthouse was receiving attestations which reference the same, un-cached shuffling on multiple threads at the same time. Lighthouse was then loading the same state from database and determining the shuffling on multiple threads at the same time. This is unnecessary load on the disk and RAM. This PR modifies the shuffling cache so that each entry can be either: - A committee - A promise for a committee (i.e., a `crossbeam_channel::Receiver`) Now, in the scenario where we have thread A and thread B simultaneously requesting the same un-cached shuffling, we will have the following: 1. Thread A will take the write-lock on the shuffling cache, find that there's no cached committee and then create a "promise" (a `crossbeam_channel::Sender`) for a committee before dropping the write-lock. 1. Thread B will then be allowed to take the write-lock for the shuffling cache and find the promise created by thread A. It will block the current thread waiting for thread A to fulfill that promise. 1. Thread A will load the state from disk, obtain the shuffling, send it down the channel, insert the entry into the cache and then continue to verify the attestation. 1. Thread B will then receive the shuffling from the receiver, be un-blocked and then continue to verify the attestation. In the case where thread A fails to generate the shuffling and drops the sender, the next time that specific shuffling is requested we will detect that the channel is disconnected and return a `None` entry for that shuffling. This will cause the shuffling to be re-calculated. ## Additional Info NA
2022-09-16 08:54:03 +00:00
.insert_committee_cache(shuffling_id, committee_cache);
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
}
Optimize attestation processing (#841) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Disable/delete two outdated tests * Tidy * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Tidy * Fix todos * Add attestation processing tests * Add another test * Only run attestation tests in release * Make attestation tests MainnetEthSpec * Address Michael's comments * Remove redundant check * Fix warning * Fix failing test Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-03-05 06:19:35 +00:00
}
// Apply the state to the attester cache, only if it is from the previous epoch or later.
//
// In a perfect scenario there should be no need to add previous-epoch states to the cache.
// However, latency between the VC and the BN might cause the VC to produce attestations at
// a previous slot.
if state.current_epoch().saturating_add(1_u64) >= current_epoch {
self.attester_cache
.maybe_cache_state(&state, block_root, &self.spec)
.map_err(BeaconChainError::from)?;
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// Alias for readability.
let block = signed_block.message();
// Only perform the weak subjectivity check if it was configured.
if let Some(wss_checkpoint) = self.config.weak_subjectivity_checkpoint {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// Note: we're using the finalized checkpoint from the head state, rather than fork
// choice.
//
// We are doing this to ensure that we detect changes in finalization. It's possible
// that fork choice has already been updated to the finalized checkpoint in the block
// we're importing.
let current_head_finalized_checkpoint =
self.canonical_head.cached_head().finalized_checkpoint();
// Compare the existing finalized checkpoint with the incoming block's finalized checkpoint.
let new_finalized_checkpoint = state.finalized_checkpoint();
// This ensures we only perform the check once.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
if (current_head_finalized_checkpoint.epoch < wss_checkpoint.epoch)
&& (wss_checkpoint.epoch <= new_finalized_checkpoint.epoch)
{
if let Err(e) =
self.verify_weak_subjectivity_checkpoint(wss_checkpoint, block_root, &state)
{
let mut shutdown_sender = self.shutdown_sender();
crit!(
self.log,
"Weak subjectivity checkpoint verification failed while importing block!";
"block_root" => ?block_root,
"parent_root" => ?block.parent_root(),
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
"old_finalized_epoch" => ?current_head_finalized_checkpoint.epoch,
"new_finalized_epoch" => ?new_finalized_checkpoint.epoch,
"weak_subjectivity_epoch" => ?wss_checkpoint.epoch,
"error" => ?e,
);
crit!(self.log, "You must use the `--purge-db` flag to clear the database and restart sync. You may be on a hostile network.");
shutdown_sender
.try_send(ShutdownReason::Failure(
"Weak subjectivity checkpoint verification failed. Provided block root is not a checkpoint."
))
.map_err(|err| BlockError::BeaconChainError(BeaconChainError::WeakSubjectivtyShutdownError(err)))?;
return Err(BlockError::WeakSubjectivityConflict);
}
}
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// Take an exclusive write-lock on fork choice. It's very important prevent deadlocks by
// avoiding taking other locks whilst holding this lock.
let mut fork_choice = self.canonical_head.fork_choice_write_lock();
// Do not import a block that doesn't descend from the finalized root.
check_block_is_finalized_descendant(self, &fork_choice, &signed_block)?;
Add proto_array fork choice (#804) * Start implementing proto_array * Add progress * Add unfinished progress * Add further progress * Add progress * Add tree filtering * Add half-finished modifications * Add refactored version * Tidy, add incomplete LmdGhost impl * Move impls in LmdGhost trait def * Remove old reduced_tree fork choice * Combine two functions in to `compute_deltas` * Start testing * Add more compute_deltas tests * Add fork choice testing * Add more fork choice testing * Add more fork choice tests * Add more testing to proto-array * Remove old tests * Modify tests * Add more tests * Add more testing * Add comments and fixes * Re-organise crate * Tidy, finish pruning tests * Add ssz encoding, other pub fns * Rename lmd_ghost > proto_array_fork_choice * Integrate proto_array into lighthouse * Add first pass at fixing filter * Clean out old comments * Add more comments * Attempt to fix prune error * Adjust TODO * Fix test compile errors * Add extra justification change check * Update cargo.lock * Fix fork choice test compile errors * Most remove ffg_update_required * Fix bug with epoch of attestation votes * Start adding new test format * Make fork choice tests declarative * Create test def concept * Move test defs into crate * Add binary, re-org crate * Shuffle files * Start adding ffg tests * Add more fork choice tests * Add fork choice JSON dumping * Add more detail to best node error * Ensure fin+just checkpoints from from same block * Rename JustificationManager * Move checkpoint manager into own file * Tidy * Add targetted logging for sneaky sync bug * Fix justified balances bug * Add cache metrics * Add metrics for log levels * Fix bug in checkpoint manager * Fix compile error in fork choice tests * Ignore duplicate blocks in fork choice * Add block to fock choice before db * Rename on_new_block fn * Fix spec inconsistency in `CheckpointManager` * Remove BlockRootTree * Remove old reduced_tree code fragment * Add API endpoint for fork choice * Add more ffg tests * Remove block_root_tree reminents * Ensure effective balances are used * Remove old debugging code, fix API fault * Add check to ensure parent block is in fork choice * Update readme dates * Fix readme * Tidy checkpoint manager * Remove fork choice yaml files from repo * Remove fork choice yaml from repo * General tidy * Address majority of Michael's comments * Tidy bin/lib business * Remove dangling file * Undo changes for rpc/handler from master * Revert "Undo changes for rpc/handler from master" This reverts commit 876edff0e4a501aafbb47113454852826dcc24e8. Co-authored-by: Age Manning <Age@AgeManning.com>
2020-01-29 04:05:00 +00:00
// Register the new block with the fork choice service.
{
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
let _fork_choice_block_timer =
metrics::start_timer(&metrics::FORK_CHOICE_PROCESS_BLOCK_TIMES);
v1.1.6 Fork Choice changes (#2822) ## Issue Addressed Resolves: https://github.com/sigp/lighthouse/issues/2741 Includes: https://github.com/sigp/lighthouse/pull/2853 so that we can get ssz static tests passing here on v1.1.6. If we want to merge that first, we can make this diff slightly smaller ## Proposed Changes - Changes the `justified_epoch` and `finalized_epoch` in the `ProtoArrayNode` each to an `Option<Checkpoint>`. The `Option` is necessary only for the migration, so not ideal. But does allow us to add a default logic to `None` on these fields during the database migration. - Adds a database migration from a legacy fork choice struct to the new one, search for all necessary block roots in fork choice by iterating through blocks in the db. - updates related to https://github.com/ethereum/consensus-specs/pull/2727 - We will have to update the persisted forkchoice to make sure the justified checkpoint stored is correct according to the updated fork choice logic. This boils down to setting the forkchoice store's justified checkpoint to the justified checkpoint of the block that advanced the finalized checkpoint to the current one. - AFAICT there's no migration steps necessary for the update to allow applying attestations from prior blocks, but would appreciate confirmation on that - I updated the consensus spec tests to v1.1.6 here, but they will fail until we also implement the proposer score boost updates. I confirmed that the previously failing scenario `new_finalized_slot_is_justified_checkpoint_ancestor` will now pass after the boost updates, but haven't confirmed _all_ tests will pass because I just quickly stubbed out the proposer boost test scenario formatting. - This PR now also includes proposer boosting https://github.com/ethereum/consensus-specs/pull/2730 ## Additional Info I realized checking justified and finalized roots in fork choice makes it more likely that we trigger this bug: https://github.com/ethereum/consensus-specs/pull/2727 It's possible the combination of justified checkpoint and finalized checkpoint in the forkchoice store is different from in any block in fork choice. So when trying to startup our store's justified checkpoint seems invalid to the rest of fork choice (but it should be valid). When this happens we get an `InvalidBestNode` error and fail to start up. So I'm including that bugfix in this branch. Todo: - [x] Fix fork choice tests - [x] Self review - [x] Add fix for https://github.com/ethereum/consensus-specs/pull/2727 - [x] Rebase onto Kintusgi - [x] Fix `num_active_validators` calculation as @michaelsproul pointed out - [x] Clean up db migrations Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-12-13 20:43:22 +00:00
let block_delay = self
.slot_clock
.seconds_from_current_slot_start(self.spec.seconds_per_slot)
.ok_or(Error::UnableToComputeTimeAtSlot)?;
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
fork_choice
.on_block(
current_slot,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
block,
block_root,
v1.1.6 Fork Choice changes (#2822) ## Issue Addressed Resolves: https://github.com/sigp/lighthouse/issues/2741 Includes: https://github.com/sigp/lighthouse/pull/2853 so that we can get ssz static tests passing here on v1.1.6. If we want to merge that first, we can make this diff slightly smaller ## Proposed Changes - Changes the `justified_epoch` and `finalized_epoch` in the `ProtoArrayNode` each to an `Option<Checkpoint>`. The `Option` is necessary only for the migration, so not ideal. But does allow us to add a default logic to `None` on these fields during the database migration. - Adds a database migration from a legacy fork choice struct to the new one, search for all necessary block roots in fork choice by iterating through blocks in the db. - updates related to https://github.com/ethereum/consensus-specs/pull/2727 - We will have to update the persisted forkchoice to make sure the justified checkpoint stored is correct according to the updated fork choice logic. This boils down to setting the forkchoice store's justified checkpoint to the justified checkpoint of the block that advanced the finalized checkpoint to the current one. - AFAICT there's no migration steps necessary for the update to allow applying attestations from prior blocks, but would appreciate confirmation on that - I updated the consensus spec tests to v1.1.6 here, but they will fail until we also implement the proposer score boost updates. I confirmed that the previously failing scenario `new_finalized_slot_is_justified_checkpoint_ancestor` will now pass after the boost updates, but haven't confirmed _all_ tests will pass because I just quickly stubbed out the proposer boost test scenario formatting. - This PR now also includes proposer boosting https://github.com/ethereum/consensus-specs/pull/2730 ## Additional Info I realized checking justified and finalized roots in fork choice makes it more likely that we trigger this bug: https://github.com/ethereum/consensus-specs/pull/2727 It's possible the combination of justified checkpoint and finalized checkpoint in the forkchoice store is different from in any block in fork choice. So when trying to startup our store's justified checkpoint seems invalid to the rest of fork choice (but it should be valid). When this happens we get an `InvalidBestNode` error and fail to start up. So I'm including that bugfix in this branch. Todo: - [x] Fix fork choice tests - [x] Self review - [x] Add fix for https://github.com/ethereum/consensus-specs/pull/2727 - [x] Rebase onto Kintusgi - [x] Fix `num_active_validators` calculation as @michaelsproul pointed out - [x] Clean up db migrations Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-12-13 20:43:22 +00:00
block_delay,
&state,
payload_verification_status,
&self.spec,
count_unrealized.and(self.config.count_unrealized.into()),
)
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
.map_err(|e| BlockError::BeaconChainError(e.into()))?;
Add proto_array fork choice (#804) * Start implementing proto_array * Add progress * Add unfinished progress * Add further progress * Add progress * Add tree filtering * Add half-finished modifications * Add refactored version * Tidy, add incomplete LmdGhost impl * Move impls in LmdGhost trait def * Remove old reduced_tree fork choice * Combine two functions in to `compute_deltas` * Start testing * Add more compute_deltas tests * Add fork choice testing * Add more fork choice testing * Add more fork choice tests * Add more testing to proto-array * Remove old tests * Modify tests * Add more tests * Add more testing * Add comments and fixes * Re-organise crate * Tidy, finish pruning tests * Add ssz encoding, other pub fns * Rename lmd_ghost > proto_array_fork_choice * Integrate proto_array into lighthouse * Add first pass at fixing filter * Clean out old comments * Add more comments * Attempt to fix prune error * Adjust TODO * Fix test compile errors * Add extra justification change check * Update cargo.lock * Fix fork choice test compile errors * Most remove ffg_update_required * Fix bug with epoch of attestation votes * Start adding new test format * Make fork choice tests declarative * Create test def concept * Move test defs into crate * Add binary, re-org crate * Shuffle files * Start adding ffg tests * Add more fork choice tests * Add fork choice JSON dumping * Add more detail to best node error * Ensure fin+just checkpoints from from same block * Rename JustificationManager * Move checkpoint manager into own file * Tidy * Add targetted logging for sneaky sync bug * Fix justified balances bug * Add cache metrics * Add metrics for log levels * Fix bug in checkpoint manager * Fix compile error in fork choice tests * Ignore duplicate blocks in fork choice * Add block to fock choice before db * Rename on_new_block fn * Fix spec inconsistency in `CheckpointManager` * Remove BlockRootTree * Remove old reduced_tree code fragment * Add API endpoint for fork choice * Add more ffg tests * Remove block_root_tree reminents * Ensure effective balances are used * Remove old debugging code, fix API fault * Add check to ensure parent block is in fork choice * Update readme dates * Fix readme * Tidy checkpoint manager * Remove fork choice yaml files from repo * Remove fork choice yaml from repo * General tidy * Address majority of Michael's comments * Tidy bin/lib business * Remove dangling file * Undo changes for rpc/handler from master * Revert "Undo changes for rpc/handler from master" This reverts commit 876edff0e4a501aafbb47113454852826dcc24e8. Co-authored-by: Age Manning <Age@AgeManning.com>
2020-01-29 04:05:00 +00:00
}
// Allow the validator monitor to learn about a new valid state.
self.validator_monitor
.write()
.process_valid_state(current_slot.epoch(T::EthSpec::slots_per_epoch()), &state);
let validator_monitor = self.validator_monitor.read();
// Register each attester slashing in the block with fork choice.
for attester_slashing in block.body().attester_slashings() {
fork_choice.on_attester_slashing(attester_slashing);
}
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
// Register each attestation in the block with the fork choice service.
for attestation in block.body().attestations() {
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
let _fork_choice_attestation_timer =
metrics::start_timer(&metrics::FORK_CHOICE_PROCESS_ATTESTATION_TIMES);
Ensure doppelganger detects attestations in blocks (#2495) ## Issue Addressed NA ## Proposed Changes When testing our (not-yet-released) Doppelganger implementation, I noticed that we aren't detecting attestations included in blocks (only those on the gossip network). This is because during [block processing](https://github.com/sigp/lighthouse/blob/e8c0d1f19b2736efb83c67a247e0022da5eaa7bb/beacon_node/beacon_chain/src/beacon_chain.rs#L2168) we only update the `observed_attestations` cache with each attestation, but not the `observed_attesters` cache. This is the correct behaviour when we consider the [p2p spec](https://github.com/ethereum/eth2.0-specs/blob/v1.0.1/specs/phase0/p2p-interface.md): > [IGNORE] There has been no other valid attestation seen on an attestation subnet that has an identical attestation.data.target.epoch and participating validator index. We're doing the right thing here and still allowing attestations on gossip that we've seen in a block. However, this doesn't work so nicely for Doppelganger. To resolve this, I've taken the following steps: - Add a `observed_block_attesters` cache. - Rename `observed_attesters` to `observed_gossip_attesters`. ## TODO - [x] Add a test to ensure a validator that's been seen in a block attestation (but not a gossip attestation) returns `true` for `BeaconChain::validator_seen_at_epoch`. - [x] Add a test to ensure `observed_block_attesters` isn't polluted via gossip attestations and vice versa. Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-08-09 02:43:03 +00:00
let attestation_target_epoch = attestation.data.target.epoch;
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
let committee =
state.get_beacon_committee(attestation.data.slot, attestation.data.index)?;
let indexed_attestation = get_indexed_attestation(committee.committee, attestation)
.map_err(|e| BlockError::BeaconChainError(e.into()))?;
v1.1.6 Fork Choice changes (#2822) ## Issue Addressed Resolves: https://github.com/sigp/lighthouse/issues/2741 Includes: https://github.com/sigp/lighthouse/pull/2853 so that we can get ssz static tests passing here on v1.1.6. If we want to merge that first, we can make this diff slightly smaller ## Proposed Changes - Changes the `justified_epoch` and `finalized_epoch` in the `ProtoArrayNode` each to an `Option<Checkpoint>`. The `Option` is necessary only for the migration, so not ideal. But does allow us to add a default logic to `None` on these fields during the database migration. - Adds a database migration from a legacy fork choice struct to the new one, search for all necessary block roots in fork choice by iterating through blocks in the db. - updates related to https://github.com/ethereum/consensus-specs/pull/2727 - We will have to update the persisted forkchoice to make sure the justified checkpoint stored is correct according to the updated fork choice logic. This boils down to setting the forkchoice store's justified checkpoint to the justified checkpoint of the block that advanced the finalized checkpoint to the current one. - AFAICT there's no migration steps necessary for the update to allow applying attestations from prior blocks, but would appreciate confirmation on that - I updated the consensus spec tests to v1.1.6 here, but they will fail until we also implement the proposer score boost updates. I confirmed that the previously failing scenario `new_finalized_slot_is_justified_checkpoint_ancestor` will now pass after the boost updates, but haven't confirmed _all_ tests will pass because I just quickly stubbed out the proposer boost test scenario formatting. - This PR now also includes proposer boosting https://github.com/ethereum/consensus-specs/pull/2730 ## Additional Info I realized checking justified and finalized roots in fork choice makes it more likely that we trigger this bug: https://github.com/ethereum/consensus-specs/pull/2727 It's possible the combination of justified checkpoint and finalized checkpoint in the forkchoice store is different from in any block in fork choice. So when trying to startup our store's justified checkpoint seems invalid to the rest of fork choice (but it should be valid). When this happens we get an `InvalidBestNode` error and fail to start up. So I'm including that bugfix in this branch. Todo: - [x] Fix fork choice tests - [x] Self review - [x] Add fix for https://github.com/ethereum/consensus-specs/pull/2727 - [x] Rebase onto Kintusgi - [x] Fix `num_active_validators` calculation as @michaelsproul pointed out - [x] Clean up db migrations Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-12-13 20:43:22 +00:00
match fork_choice.on_attestation(
current_slot,
&indexed_attestation,
AttestationFromBlock::True,
&self.spec,
v1.1.6 Fork Choice changes (#2822) ## Issue Addressed Resolves: https://github.com/sigp/lighthouse/issues/2741 Includes: https://github.com/sigp/lighthouse/pull/2853 so that we can get ssz static tests passing here on v1.1.6. If we want to merge that first, we can make this diff slightly smaller ## Proposed Changes - Changes the `justified_epoch` and `finalized_epoch` in the `ProtoArrayNode` each to an `Option<Checkpoint>`. The `Option` is necessary only for the migration, so not ideal. But does allow us to add a default logic to `None` on these fields during the database migration. - Adds a database migration from a legacy fork choice struct to the new one, search for all necessary block roots in fork choice by iterating through blocks in the db. - updates related to https://github.com/ethereum/consensus-specs/pull/2727 - We will have to update the persisted forkchoice to make sure the justified checkpoint stored is correct according to the updated fork choice logic. This boils down to setting the forkchoice store's justified checkpoint to the justified checkpoint of the block that advanced the finalized checkpoint to the current one. - AFAICT there's no migration steps necessary for the update to allow applying attestations from prior blocks, but would appreciate confirmation on that - I updated the consensus spec tests to v1.1.6 here, but they will fail until we also implement the proposer score boost updates. I confirmed that the previously failing scenario `new_finalized_slot_is_justified_checkpoint_ancestor` will now pass after the boost updates, but haven't confirmed _all_ tests will pass because I just quickly stubbed out the proposer boost test scenario formatting. - This PR now also includes proposer boosting https://github.com/ethereum/consensus-specs/pull/2730 ## Additional Info I realized checking justified and finalized roots in fork choice makes it more likely that we trigger this bug: https://github.com/ethereum/consensus-specs/pull/2727 It's possible the combination of justified checkpoint and finalized checkpoint in the forkchoice store is different from in any block in fork choice. So when trying to startup our store's justified checkpoint seems invalid to the rest of fork choice (but it should be valid). When this happens we get an `InvalidBestNode` error and fail to start up. So I'm including that bugfix in this branch. Todo: - [x] Fix fork choice tests - [x] Self review - [x] Add fix for https://github.com/ethereum/consensus-specs/pull/2727 - [x] Rebase onto Kintusgi - [x] Fix `num_active_validators` calculation as @michaelsproul pointed out - [x] Clean up db migrations Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-12-13 20:43:22 +00:00
) {
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
Ok(()) => Ok(()),
// Ignore invalid attestations whilst importing attestations from a block. The
// block might be very old and therefore the attestations useless to fork choice.
Err(ForkChoiceError::InvalidAttestation(_)) => Ok(()),
Err(e) => Err(BlockError::BeaconChainError(e.into())),
}?;
Ensure doppelganger detects attestations in blocks (#2495) ## Issue Addressed NA ## Proposed Changes When testing our (not-yet-released) Doppelganger implementation, I noticed that we aren't detecting attestations included in blocks (only those on the gossip network). This is because during [block processing](https://github.com/sigp/lighthouse/blob/e8c0d1f19b2736efb83c67a247e0022da5eaa7bb/beacon_node/beacon_chain/src/beacon_chain.rs#L2168) we only update the `observed_attestations` cache with each attestation, but not the `observed_attesters` cache. This is the correct behaviour when we consider the [p2p spec](https://github.com/ethereum/eth2.0-specs/blob/v1.0.1/specs/phase0/p2p-interface.md): > [IGNORE] There has been no other valid attestation seen on an attestation subnet that has an identical attestation.data.target.epoch and participating validator index. We're doing the right thing here and still allowing attestations on gossip that we've seen in a block. However, this doesn't work so nicely for Doppelganger. To resolve this, I've taken the following steps: - Add a `observed_block_attesters` cache. - Rename `observed_attesters` to `observed_gossip_attesters`. ## TODO - [x] Add a test to ensure a validator that's been seen in a block attestation (but not a gossip attestation) returns `true` for `BeaconChain::validator_seen_at_epoch`. - [x] Add a test to ensure `observed_block_attesters` isn't polluted via gossip attestations and vice versa. Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-08-09 02:43:03 +00:00
// To avoid slowing down sync, only register attestations for the
// `observed_block_attesters` if they are from the previous epoch or later.
if attestation_target_epoch + 1 >= current_epoch {
let mut observed_block_attesters = self.observed_block_attesters.write();
for &validator_index in &indexed_attestation.attesting_indices {
if let Err(e) = observed_block_attesters
.observe_validator(attestation_target_epoch, validator_index as usize)
{
debug!(
self.log,
"Failed to register observed block attester";
"error" => ?e,
"epoch" => attestation_target_epoch,
"validator_index" => validator_index,
)
}
}
}
// Only register this with the validator monitor when the block is sufficiently close to
// the current slot.
if VALIDATOR_MONITOR_HISTORIC_EPOCHS as u64 * T::EthSpec::slots_per_epoch()
+ block.slot().as_u64()
>= current_slot.as_u64()
{
Improve valmon inclusion delay calculation (#2618) ## Issue Addressed Resolves #2552 ## Proposed Changes Offers some improvement in inclusion distance calculation in the validator monitor. When registering an attestation from a block, instead of doing `block.slot() - attesstation.data.slot()` to get the inclusion distance, we now pass the parent block slot from the beacon chain and do `parent_slot.saturating_sub(attestation.data.slot())`. This allows us to give best effort inclusion distance in scenarios where the attestation was included right after a skip slot. Note that this does not give accurate results in scenarios where the attestation was included few blocks after the skip slot. In this case, if the attestation slot was `b1` and was included in block `b2` with a skip slot in between, we would get the inclusion delay as 0 (by ignoring the skip slot) which is the best effort inclusion delay. ``` b1 <- missed <- b2 ``` Here, if the attestation slot was `b1` and was included in block `b3` with a skip slot and valid block `b2` in between, then we would get the inclusion delay as 2 instead of 1 (by ignoring the skip slot). ``` b1 <- missed <- b2 <- b3 ``` A solution for the scenario 2 would be to count number of slots between included slot and attestation slot ignoring the skip slots in the beacon chain and pass the value to the validator monitor. But I'm concerned that it could potentially lead to db accesses for older blocks in extreme cases. This PR also uses the validator monitor data for logging per epoch inclusion distance. This is useful as we won't get inclusion data in post-altair summaries. Co-authored-by: Michael Sproul <micsproul@gmail.com>
2021-09-30 01:22:43 +00:00
match fork_choice.get_block(&block.parent_root()) {
Some(parent_block) => validator_monitor.register_attestation_in_block(
&indexed_attestation,
parent_block.slot,
&self.spec,
),
None => warn!(self.log, "Failed to get parent block"; "slot" => %block.slot()),
}
}
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
}
Add proto_array fork choice (#804) * Start implementing proto_array * Add progress * Add unfinished progress * Add further progress * Add progress * Add tree filtering * Add half-finished modifications * Add refactored version * Tidy, add incomplete LmdGhost impl * Move impls in LmdGhost trait def * Remove old reduced_tree fork choice * Combine two functions in to `compute_deltas` * Start testing * Add more compute_deltas tests * Add fork choice testing * Add more fork choice testing * Add more fork choice tests * Add more testing to proto-array * Remove old tests * Modify tests * Add more tests * Add more testing * Add comments and fixes * Re-organise crate * Tidy, finish pruning tests * Add ssz encoding, other pub fns * Rename lmd_ghost > proto_array_fork_choice * Integrate proto_array into lighthouse * Add first pass at fixing filter * Clean out old comments * Add more comments * Attempt to fix prune error * Adjust TODO * Fix test compile errors * Add extra justification change check * Update cargo.lock * Fix fork choice test compile errors * Most remove ffg_update_required * Fix bug with epoch of attestation votes * Start adding new test format * Make fork choice tests declarative * Create test def concept * Move test defs into crate * Add binary, re-org crate * Shuffle files * Start adding ffg tests * Add more fork choice tests * Add fork choice JSON dumping * Add more detail to best node error * Ensure fin+just checkpoints from from same block * Rename JustificationManager * Move checkpoint manager into own file * Tidy * Add targetted logging for sneaky sync bug * Fix justified balances bug * Add cache metrics * Add metrics for log levels * Fix bug in checkpoint manager * Fix compile error in fork choice tests * Ignore duplicate blocks in fork choice * Add block to fock choice before db * Rename on_new_block fn * Fix spec inconsistency in `CheckpointManager` * Remove BlockRootTree * Remove old reduced_tree code fragment * Add API endpoint for fork choice * Add more ffg tests * Remove block_root_tree reminents * Ensure effective balances are used * Remove old debugging code, fix API fault * Add check to ensure parent block is in fork choice * Update readme dates * Fix readme * Tidy checkpoint manager * Remove fork choice yaml files from repo * Remove fork choice yaml from repo * General tidy * Address majority of Michael's comments * Tidy bin/lib business * Remove dangling file * Undo changes for rpc/handler from master * Revert "Undo changes for rpc/handler from master" This reverts commit 876edff0e4a501aafbb47113454852826dcc24e8. Co-authored-by: Age Manning <Age@AgeManning.com>
2020-01-29 04:05:00 +00:00
// If the block is recent enough and it was not optimistically imported, check to see if it
// becomes the head block. If so, apply it to the early attester cache. This will allow
// attestations to the block without waiting for the block and state to be inserted to the
// database.
Add early attester cache (#2872) ## Issue Addressed NA ## Proposed Changes Introduces a cache to attestation to produce atop blocks which will become the head, but are not fully imported (e.g., not inserted into the database). Whilst attesting to a block before it's imported is rather easy, if we're going to produce that attestation then we also need to be able to: 1. Verify that attestation. 1. Respond to RPC requests for the `beacon_block_root`. Attestation verification (1) is *partially* covered. Since we prime the shuffling cache before we insert the block into the early attester cache, we should be fine for all typical use-cases. However, it is possible that the cache is washed out before we've managed to insert the state into the database and then attestation verification will fail with a "missing beacon state"-type error. Providing the block via RPC (2) is also partially covered, since we'll check the database *and* the early attester cache when responding a blocks-by-root request. However, we'll still omit the block from blocks-by-range requests (until the block lands in the DB). I *think* this is fine, since there's no guarantee that we return all blocks for those responses. Another important consideration is whether or not the *parent* of the early attester block is available in the databse. If it were not, we might fail to respond to blocks-by-root request that are iterating backwards to collect a chain of blocks. I argue that *we will always have the parent of the early attester block in the database.* This is because we are holding the fork-choice write-lock when inserting the block into the early attester cache and we do not drop that until the block is in the database.
2022-01-11 01:35:55 +00:00
//
// Only performing this check on recent blocks avoids slowing down sync with lots of calls
// to fork choice `get_head`.
//
// Optimistically imported blocks are not added to the cache since the cache is only useful
// for a small window of time and the complexity of keeping track of the optimistic status
// is not worth it.
if !payload_verification_status.is_optimistic()
&& block.slot() + EARLY_ATTESTER_CACHE_HISTORIC_SLOTS >= current_slot
{
match fork_choice.get_head(current_slot, &self.spec) {
// This block became the head, add it to the early attester cache.
Ok(new_head_root) if new_head_root == block_root => {
if let Some(proto_block) = fork_choice.get_block(&block_root) {
if let Err(e) = self.early_attester_cache.add_head_block(
block_root,
signed_block.clone(),
proto_block,
&state,
&self.spec,
) {
warn!(
self.log,
"Early attester cache insert failed";
"error" => ?e
);
}
} else {
Add early attester cache (#2872) ## Issue Addressed NA ## Proposed Changes Introduces a cache to attestation to produce atop blocks which will become the head, but are not fully imported (e.g., not inserted into the database). Whilst attesting to a block before it's imported is rather easy, if we're going to produce that attestation then we also need to be able to: 1. Verify that attestation. 1. Respond to RPC requests for the `beacon_block_root`. Attestation verification (1) is *partially* covered. Since we prime the shuffling cache before we insert the block into the early attester cache, we should be fine for all typical use-cases. However, it is possible that the cache is washed out before we've managed to insert the state into the database and then attestation verification will fail with a "missing beacon state"-type error. Providing the block via RPC (2) is also partially covered, since we'll check the database *and* the early attester cache when responding a blocks-by-root request. However, we'll still omit the block from blocks-by-range requests (until the block lands in the DB). I *think* this is fine, since there's no guarantee that we return all blocks for those responses. Another important consideration is whether or not the *parent* of the early attester block is available in the databse. If it were not, we might fail to respond to blocks-by-root request that are iterating backwards to collect a chain of blocks. I argue that *we will always have the parent of the early attester block in the database.* This is because we are holding the fork-choice write-lock when inserting the block into the early attester cache and we do not drop that until the block is in the database.
2022-01-11 01:35:55 +00:00
warn!(
self.log,
"Early attester block missing";
"block_root" => ?block_root
Add early attester cache (#2872) ## Issue Addressed NA ## Proposed Changes Introduces a cache to attestation to produce atop blocks which will become the head, but are not fully imported (e.g., not inserted into the database). Whilst attesting to a block before it's imported is rather easy, if we're going to produce that attestation then we also need to be able to: 1. Verify that attestation. 1. Respond to RPC requests for the `beacon_block_root`. Attestation verification (1) is *partially* covered. Since we prime the shuffling cache before we insert the block into the early attester cache, we should be fine for all typical use-cases. However, it is possible that the cache is washed out before we've managed to insert the state into the database and then attestation verification will fail with a "missing beacon state"-type error. Providing the block via RPC (2) is also partially covered, since we'll check the database *and* the early attester cache when responding a blocks-by-root request. However, we'll still omit the block from blocks-by-range requests (until the block lands in the DB). I *think* this is fine, since there's no guarantee that we return all blocks for those responses. Another important consideration is whether or not the *parent* of the early attester block is available in the databse. If it were not, we might fail to respond to blocks-by-root request that are iterating backwards to collect a chain of blocks. I argue that *we will always have the parent of the early attester block in the database.* This is because we are holding the fork-choice write-lock when inserting the block into the early attester cache and we do not drop that until the block is in the database.
2022-01-11 01:35:55 +00:00
);
}
}
// This block did not become the head, nothing to do.
Ok(_) => (),
Err(e) => error!(
self.log,
"Failed to compute head during block import";
"error" => ?e
),
Add early attester cache (#2872) ## Issue Addressed NA ## Proposed Changes Introduces a cache to attestation to produce atop blocks which will become the head, but are not fully imported (e.g., not inserted into the database). Whilst attesting to a block before it's imported is rather easy, if we're going to produce that attestation then we also need to be able to: 1. Verify that attestation. 1. Respond to RPC requests for the `beacon_block_root`. Attestation verification (1) is *partially* covered. Since we prime the shuffling cache before we insert the block into the early attester cache, we should be fine for all typical use-cases. However, it is possible that the cache is washed out before we've managed to insert the state into the database and then attestation verification will fail with a "missing beacon state"-type error. Providing the block via RPC (2) is also partially covered, since we'll check the database *and* the early attester cache when responding a blocks-by-root request. However, we'll still omit the block from blocks-by-range requests (until the block lands in the DB). I *think* this is fine, since there's no guarantee that we return all blocks for those responses. Another important consideration is whether or not the *parent* of the early attester block is available in the databse. If it were not, we might fail to respond to blocks-by-root request that are iterating backwards to collect a chain of blocks. I argue that *we will always have the parent of the early attester block in the database.* This is because we are holding the fork-choice write-lock when inserting the block into the early attester cache and we do not drop that until the block is in the database.
2022-01-11 01:35:55 +00:00
}
}
// Register sync aggregate with validator monitor
if let Ok(sync_aggregate) = block.body().sync_aggregate() {
// `SyncCommittee` for the sync_aggregate should correspond to the duty slot
let duty_epoch = block.slot().epoch(T::EthSpec::slots_per_epoch());
let sync_committee = self.sync_committee_at_epoch(duty_epoch)?;
let participant_pubkeys = sync_committee
.pubkeys
.iter()
.zip(sync_aggregate.sync_committee_bits.iter())
.filter_map(|(pubkey, bit)| bit.then_some(pubkey))
.collect::<Vec<_>>();
validator_monitor.register_sync_aggregate_in_block(
block.slot(),
block.parent_root(),
participant_pubkeys,
);
}
for exit in block.body().voluntary_exits() {
validator_monitor.register_block_voluntary_exit(&exit.message)
}
for slashing in block.body().attester_slashings() {
validator_monitor.register_block_attester_slashing(slashing)
}
for slashing in block.body().proposer_slashings() {
validator_monitor.register_block_proposer_slashing(slashing)
}
drop(validator_monitor);
// Only present some metrics for blocks from the previous epoch or later.
//
// This helps avoid noise in the metrics during sync.
if block.slot().epoch(T::EthSpec::slots_per_epoch()) + 1 >= self.epoch()? {
metrics::observe(
&metrics::OPERATIONS_PER_BLOCK_ATTESTATION,
block.body().attestations().len() as f64,
);
if let Ok(sync_aggregate) = block.body().sync_aggregate() {
metrics::set_gauge(
&metrics::BLOCK_SYNC_AGGREGATE_SET_BITS,
sync_aggregate.num_set_bits() as i64,
);
}
}
Fix block processing blowup, upgrade metrics (#500) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Begin metrics refactor * Move beacon_chain to new metrics structure. * Make metrics not panic if already defined * Use global prometheus gather at rest api * Unify common metric fns into a crate * Add heavy metering to block processing * Remove hypen from prometheus metric name * Add more beacon chain metrics * Add beacon chain persistence metric * Prune op pool on finalization * Add extra prom beacon chain metrics * Prefix BeaconChain metrics with "beacon_" * Add more store metrics * Add basic metrics to libp2p * Add metrics to HTTP server * Remove old `http_server` crate * Update metrics names to be more like standard * Fix broken beacon chain metrics, add slot clock metrics * Add lighthouse_metrics gather fn * Remove http args * Fix wrong state given to op pool prune * Make prom metric names more consistent * Add more metrics, tidy existing metrics * Fix store block read metrics * Tidy attestation metrics * Fix minor PR comments * Allow travis failures on beta (see desc) There's a non-backward compatible change in `cargo fmt`. Stable and beta do not agree. * Tidy `lighthouse_metrics` docs * Fix typo
2019-08-19 11:02:34 +00:00
let db_write_timer = metrics::start_timer(&metrics::BLOCK_PROCESSING_DB_WRITE);
Implement database temp states to reduce memory usage (#1798) ## Issue Addressed Closes #800 Closes #1713 ## Proposed Changes Implement the temporary state storage algorithm described in #800. Specifically: * Add `DBColumn::BeaconStateTemporary`, for storing 0-length temporary marker values. * Store intermediate states immediately as they are created, marked temporary. Delete the temporary flag if the block is processed successfully. * Add a garbage collection process to delete leftover temporary states on start-up. * Bump the database schema version to 2 so that a DB with temporary states can't accidentally be used with older versions of the software. The auto-migration is a no-op, but puts in place some infra that we can use for future migrations (e.g. #1784) ## Additional Info There are two known race conditions, one potentially causing permanent faults (hopefully rare), and the other insignificant. ### Race 1: Permanent state marked temporary EDIT: this has been fixed by the addition of a lock around the relevant critical section There are 2 threads that are trying to store 2 different blocks that share some intermediate states (e.g. they both skip some slots from the current head). Consider this sequence of events: 1. Thread 1 checks if state `s` already exists, and seeing that it doesn't, prepares an atomic commit of `(s, s_temporary_flag)`. 2. Thread 2 does the same, but also gets as far as committing the state txn, finishing the processing of its block, and _deleting_ the temporary flag. 3. Thread 1 is (finally) scheduled again, and marks `s` as temporary with its transaction. 4. a) The process is killed, or thread 1's block fails verification and the temp flag is not deleted. This is a permanent failure! Any attempt to load state `s` will fail... hope it isn't on the main chain! Alternatively (4b) happens... b) Thread 1 finishes, and re-deletes the temporary flag. In this case the failure is transient, state `s` will disappear temporarily, but will come back once thread 1 finishes running. I _hope_ that steps 1-3 only happen very rarely, and 4a even more rarely. It's hard to know This once again begs the question of why we're using LevelDB (#483), when it clearly doesn't care about atomicity! A ham-fisted fix would be to wrap the hot and cold DBs in locks, which would bring us closer to how other DBs handle read-write transactions. E.g. [LMDB only allows one R/W transaction at a time](https://docs.rs/lmdb/0.8.0/lmdb/struct.Environment.html#method.begin_rw_txn). ### Race 2: Temporary state returned from `get_state` I don't think this race really matters, but in `load_hot_state`, if another thread stores a state between when we call `load_state_temporary_flag` and when we call `load_hot_state_summary`, then we could end up returning that state even though it's only a temporary state. I can't think of any case where this would be relevant, and I suspect if it did come up, it would be safe/recoverable (having data is safer than _not_ having data). This could be fixed by using a LevelDB read snapshot, but that would require substantial changes to how we read all our values, so I don't think it's worth it right now.
2020-10-23 01:27:51 +00:00
// Store the block and its state, and execute the confirmation batch for the intermediate
// states, which will delete their temporary flags.
// If the write fails, revert fork choice to the version from disk, else we can
// end up with blocks in fork choice that are missing from disk.
// See https://github.com/sigp/lighthouse/issues/2028
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let mut ops: Vec<_> = confirmed_state_roots
.into_iter()
.map(StoreOp::DeleteStateTemporaryFlag)
.collect();
ops.push(StoreOp::PutBlock(block_root, signed_block.clone()));
ops.push(StoreOp::PutState(block.state_root(), &state));
Implement database temp states to reduce memory usage (#1798) ## Issue Addressed Closes #800 Closes #1713 ## Proposed Changes Implement the temporary state storage algorithm described in #800. Specifically: * Add `DBColumn::BeaconStateTemporary`, for storing 0-length temporary marker values. * Store intermediate states immediately as they are created, marked temporary. Delete the temporary flag if the block is processed successfully. * Add a garbage collection process to delete leftover temporary states on start-up. * Bump the database schema version to 2 so that a DB with temporary states can't accidentally be used with older versions of the software. The auto-migration is a no-op, but puts in place some infra that we can use for future migrations (e.g. #1784) ## Additional Info There are two known race conditions, one potentially causing permanent faults (hopefully rare), and the other insignificant. ### Race 1: Permanent state marked temporary EDIT: this has been fixed by the addition of a lock around the relevant critical section There are 2 threads that are trying to store 2 different blocks that share some intermediate states (e.g. they both skip some slots from the current head). Consider this sequence of events: 1. Thread 1 checks if state `s` already exists, and seeing that it doesn't, prepares an atomic commit of `(s, s_temporary_flag)`. 2. Thread 2 does the same, but also gets as far as committing the state txn, finishing the processing of its block, and _deleting_ the temporary flag. 3. Thread 1 is (finally) scheduled again, and marks `s` as temporary with its transaction. 4. a) The process is killed, or thread 1's block fails verification and the temp flag is not deleted. This is a permanent failure! Any attempt to load state `s` will fail... hope it isn't on the main chain! Alternatively (4b) happens... b) Thread 1 finishes, and re-deletes the temporary flag. In this case the failure is transient, state `s` will disappear temporarily, but will come back once thread 1 finishes running. I _hope_ that steps 1-3 only happen very rarely, and 4a even more rarely. It's hard to know This once again begs the question of why we're using LevelDB (#483), when it clearly doesn't care about atomicity! A ham-fisted fix would be to wrap the hot and cold DBs in locks, which would bring us closer to how other DBs handle read-write transactions. E.g. [LMDB only allows one R/W transaction at a time](https://docs.rs/lmdb/0.8.0/lmdb/struct.Environment.html#method.begin_rw_txn). ### Race 2: Temporary state returned from `get_state` I don't think this race really matters, but in `load_hot_state`, if another thread stores a state between when we call `load_state_temporary_flag` and when we call `load_hot_state_summary`, then we could end up returning that state even though it's only a temporary state. I can't think of any case where this would be relevant, and I suspect if it did come up, it would be safe/recoverable (having data is safer than _not_ having data). This could be fixed by using a LevelDB read snapshot, but that would require substantial changes to how we read all our values, so I don't think it's worth it right now.
2020-10-23 01:27:51 +00:00
let txn_lock = self.store.hot_db.begin_rw_transaction();
if let Err(e) = self.store.do_atomically(ops) {
error!(
self.log,
"Database write failed!";
"msg" => "Restoring fork choice from disk",
"error" => ?e,
);
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// Since the write failed, try to revert the canonical head back to what was stored
// in the database. This attempts to prevent inconsistency between the database and
// fork choice.
Reset payload statuses when resuming fork choice (#3498) ## Issue Addressed NA ## Proposed Changes This PR is motivated by a recent consensus failure in Geth where it returned `INVALID` for an `VALID` block. Without this PR, the only way to recover is by re-syncing Lighthouse. Whilst ELs "shouldn't have consensus failures", in reality it's something that we can expect from time to time due to the complex nature of Ethereum. Being able to recover easily will help the network recover and EL devs to troubleshoot. The risk introduced with this PR is that genuinely INVALID payloads get a "second chance" at being imported. I believe the DoS risk here is negligible since LH needs to be restarted in order to re-process the payload. Furthermore, there's no reason to think that a well-performing EL will accept a truly invalid payload the second-time-around. ## Additional Info This implementation has the following intricacies: 1. Instead of just resetting *invalid* payloads to optimistic, we'll also reset *valid* payloads. This is an artifact of our existing implementation. 1. We will only reset payload statuses when we detect an invalid payload present in `proto_array` - This helps save us from forgetting that all our blocks are valid in the "best case scenario" where there are no invalid blocks. 1. If we fail to revert the payload statuses we'll log a `CRIT` and just continue with a `proto_array` that *does not* have reverted payload statuses. - The code to revert statuses needs to deal with balances and proposer-boost, so it's a failure point. This is a defensive measure to avoid introducing new show-stopping bugs to LH.
2022-08-29 14:34:41 +00:00
if let Err(e) = self.canonical_head.restore_from_store(
fork_choice,
ResetPayloadStatuses::always_reset_conditionally(
self.config.always_reset_payload_statuses,
),
self.config.count_unrealized_full,
Reset payload statuses when resuming fork choice (#3498) ## Issue Addressed NA ## Proposed Changes This PR is motivated by a recent consensus failure in Geth where it returned `INVALID` for an `VALID` block. Without this PR, the only way to recover is by re-syncing Lighthouse. Whilst ELs "shouldn't have consensus failures", in reality it's something that we can expect from time to time due to the complex nature of Ethereum. Being able to recover easily will help the network recover and EL devs to troubleshoot. The risk introduced with this PR is that genuinely INVALID payloads get a "second chance" at being imported. I believe the DoS risk here is negligible since LH needs to be restarted in order to re-process the payload. Furthermore, there's no reason to think that a well-performing EL will accept a truly invalid payload the second-time-around. ## Additional Info This implementation has the following intricacies: 1. Instead of just resetting *invalid* payloads to optimistic, we'll also reset *valid* payloads. This is an artifact of our existing implementation. 1. We will only reset payload statuses when we detect an invalid payload present in `proto_array` - This helps save us from forgetting that all our blocks are valid in the "best case scenario" where there are no invalid blocks. 1. If we fail to revert the payload statuses we'll log a `CRIT` and just continue with a `proto_array` that *does not* have reverted payload statuses. - The code to revert statuses needs to deal with balances and proposer-boost, so it's a failure point. This is a defensive measure to avoid introducing new show-stopping bugs to LH.
2022-08-29 14:34:41 +00:00
&self.store,
&self.spec,
&self.log,
) {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
crit!(
self.log,
"No stored fork choice found to restore from";
"error" => ?e,
"warning" => "The database is likely corrupt now, consider --purge-db"
);
return Err(BlockError::BeaconChainError(e));
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
return Err(e.into());
}
Implement database temp states to reduce memory usage (#1798) ## Issue Addressed Closes #800 Closes #1713 ## Proposed Changes Implement the temporary state storage algorithm described in #800. Specifically: * Add `DBColumn::BeaconStateTemporary`, for storing 0-length temporary marker values. * Store intermediate states immediately as they are created, marked temporary. Delete the temporary flag if the block is processed successfully. * Add a garbage collection process to delete leftover temporary states on start-up. * Bump the database schema version to 2 so that a DB with temporary states can't accidentally be used with older versions of the software. The auto-migration is a no-op, but puts in place some infra that we can use for future migrations (e.g. #1784) ## Additional Info There are two known race conditions, one potentially causing permanent faults (hopefully rare), and the other insignificant. ### Race 1: Permanent state marked temporary EDIT: this has been fixed by the addition of a lock around the relevant critical section There are 2 threads that are trying to store 2 different blocks that share some intermediate states (e.g. they both skip some slots from the current head). Consider this sequence of events: 1. Thread 1 checks if state `s` already exists, and seeing that it doesn't, prepares an atomic commit of `(s, s_temporary_flag)`. 2. Thread 2 does the same, but also gets as far as committing the state txn, finishing the processing of its block, and _deleting_ the temporary flag. 3. Thread 1 is (finally) scheduled again, and marks `s` as temporary with its transaction. 4. a) The process is killed, or thread 1's block fails verification and the temp flag is not deleted. This is a permanent failure! Any attempt to load state `s` will fail... hope it isn't on the main chain! Alternatively (4b) happens... b) Thread 1 finishes, and re-deletes the temporary flag. In this case the failure is transient, state `s` will disappear temporarily, but will come back once thread 1 finishes running. I _hope_ that steps 1-3 only happen very rarely, and 4a even more rarely. It's hard to know This once again begs the question of why we're using LevelDB (#483), when it clearly doesn't care about atomicity! A ham-fisted fix would be to wrap the hot and cold DBs in locks, which would bring us closer to how other DBs handle read-write transactions. E.g. [LMDB only allows one R/W transaction at a time](https://docs.rs/lmdb/0.8.0/lmdb/struct.Environment.html#method.begin_rw_txn). ### Race 2: Temporary state returned from `get_state` I don't think this race really matters, but in `load_hot_state`, if another thread stores a state between when we call `load_state_temporary_flag` and when we call `load_hot_state_summary`, then we could end up returning that state even though it's only a temporary state. I can't think of any case where this would be relevant, and I suspect if it did come up, it would be safe/recoverable (having data is safer than _not_ having data). This could be fixed by using a LevelDB read snapshot, but that would require substantial changes to how we read all our values, so I don't think it's worth it right now.
2020-10-23 01:27:51 +00:00
drop(txn_lock);
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
v0.12 fork choice update (#1229) * Incomplete scraps * Add progress on new fork choice impl * Further progress * First complete compiling version * Remove chain reference * Add new lmd_ghost crate * Start integrating into beacon chain * Update `milagro_bls` to new release (#1183) * Update milagro_bls to new release Signed-off-by: Kirk Baird <baird.k@outlook.com> * Tidy up fake cryptos Signed-off-by: Kirk Baird <baird.k@outlook.com> * move SecretHash to bls and put plaintext back Signed-off-by: Kirk Baird <baird.k@outlook.com> * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Rough beacon chain impl working * Remove fork_choice_2 * Remove checkpoint manager * Half finished ssz impl * Add missed file * Add persistence * Tidy, fix some compile errors * Remove RwLock from ProtoArrayForkChoice * Fix store-based compile errors * Add comments, tidy * Move function out of ForkChoice struct * Start testing * More testing * Fix compile error * Tidy beacon_chain::fork_choice * Queue attestations from the current slot * Allow fork choice to handle prior-to-genesis start * Improve error granularity * Test attestation dequeuing * Process attestations during block * Store target root in fork choice * Move fork choice verification into new crate * Update tests * Consensus updates for v0.12 (#1228) * Update state processing for v0.12 * Fix EF test runners for v0.12 * Fix some tests * Fix broken attestation verification test * More test fixes * Fix typo found in review * Add `Block` struct to ProtoArray * Start fixing get_ancestor * Add rough progress on testing * Get fork choice tests working * Progress with testing * Fix partialeq impl * Move slot clock from fc_store * Improve testing * Add testing for best justified * Add clone back to SystemTimeSlotClock * Add balances test * Start adding balances cache again * Wire-in balances cache * Improve tests * Remove commented-out tests * Remove beacon_chain::ForkChoice * Rename crates * Update wider codebase to new fork_choice layout * Move advance_slot in test harness * Tidy ForkChoice::update_time * Fix verification tests * Fix compile error with iter::once * Fix fork choice tests * Ensure block attestations are processed * Fix failing beacon_chain tests * Add first invalid block check * Add finalized block check * Progress with testing, new store builder * Add fixes to get_ancestor * Fix old genesis justification test * Fix remaining fork choice tests * Change root iteration method * Move on_verified_block * Remove unused method * Start adding attestation verification tests * Add invalid ffg target test * Add target epoch test * Add queued attestation test * Remove old fork choice verification tests * Tidy, add test * Move fork choice lock drop * Rename BeaconForkChoiceStore * Add comments, tidy BeaconForkChoiceStore * Update metrics, rename fork_choice_store.rs * Remove genesis_block_root from ForkChoice * Tidy * Update fork_choice comments * Tidy, add comments * Tidy, simplify ForkChoice, fix compile issue * Tidy, removed dead file * Increase http request timeout * Fix failing rest_api test * Set HTTP timeout back to 5s * Apply fix to get_ancestor * Address Michael's comments * Fix typo * Revert "Fix broken attestation verification test" This reverts commit 722cdc903b12611de27916a57eeecfa3224f2279. Co-authored-by: Kirk Baird <baird.k@outlook.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-06-17 01:10:22 +00:00
// The fork choice write-lock is dropped *after* the on-disk database has been updated.
// This prevents inconsistency between the two at the expense of concurrency.
drop(fork_choice);
// We're declaring the block "imported" at this point, since fork choice and the DB know
// about it.
let block_time_imported = timestamp_now();
let parent_root = block.parent_root();
let slot = block.slot();
Deposit Cache Finalization & Fast WS Sync (#2915) ## Summary The deposit cache now has the ability to finalize deposits. This will cause it to drop unneeded deposit logs and hashes in the deposit Merkle tree that are no longer required to construct deposit proofs. The cache is finalized whenever the latest finalized checkpoint has a new `Eth1Data` with all deposits imported. This has three benefits: 1. Improves the speed of constructing Merkle proofs for deposits as we can just replay deposits since the last finalized checkpoint instead of all historical deposits when re-constructing the Merkle tree. 2. Significantly faster weak subjectivity sync as the deposit cache can be transferred to the newly syncing node in compressed form. The Merkle tree that stores `N` finalized deposits requires a maximum of `log2(N)` hashes. The newly syncing node then only needs to download deposits since the last finalized checkpoint to have a full tree. 3. Future proofing in preparation for [EIP-4444](https://eips.ethereum.org/EIPS/eip-4444) as execution nodes will no longer be required to store logs permanently so we won't always have all historical logs available to us. ## More Details Image to illustrate how the deposit contract merkle tree evolves and finalizes along with the resulting `DepositTreeSnapshot` ![image](https://user-images.githubusercontent.com/37123614/151465302-5fc56284-8a69-4998-b20e-45db3934ac70.png) ## Other Considerations I've changed the structure of the `SszDepositCache` so once you load & save your database from this version of lighthouse, you will no longer be able to load it from older versions. Co-authored-by: ethDreamer <37123614+ethDreamer@users.noreply.github.com>
2022-10-30 04:04:24 +00:00
let current_eth1_finalization_data = Eth1FinalizationData {
eth1_data: state.eth1_data().clone(),
eth1_deposit_index: state.eth1_deposit_index(),
};
let current_finalized_checkpoint = state.finalized_checkpoint();
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
self.snapshot_cache
.try_write_for(BLOCK_PROCESSING_CACHE_LOCK_TIMEOUT)
Advance state to next slot after importing block (#2174) ## Issue Addressed NA ## Proposed Changes Add an optimization to perform `per_slot_processing` from the *leading-edge* of block processing to the *trailing-edge*. Ultimately, this allows us to import the block at slot `n` faster because we used the tail-end of slot `n - 1` to perform `per_slot_processing`. Additionally, add a "block proposer cache" which allows us to cache the block proposer for some epoch. Since we're now doing trailing-edge `per_slot_processing`, we can prime this cache with the values for the next epoch before those blocks arrive (assuming those blocks don't have some weird forking). There were several ancillary changes required to achieve this: - Remove the `state_root` field of `BeaconSnapshot`, since there's no need to know it on a `pre_state` and in all other cases we can just read it from `block.state_root()`. - This caused some "dust" changes of `snapshot.beacon_state_root` to `snapshot.beacon_state_root()`, where the `BeaconSnapshot::beacon_state_root()` func just reads the state root from the block. - Rename `types::ShuffingId` to `AttestationShufflingId`. I originally did this because I added a `ProposerShufflingId` struct which turned out to be not so useful. I thought this new name was more descriptive so I kept it. - Address https://github.com/ethereum/eth2.0-specs/pull/2196 - Add a debug log when we get a block with an unknown parent. There was previously no logging around this case. - Add a function to `BeaconState` to compute all proposers for an epoch without re-computing the active indices for each slot. ## Additional Info - ~~Blocked on #2173~~ - ~~Blocked on #2179~~ That PR was wrapped into this PR. - There's potentially some places where we could avoid computing the proposer indices in `per_block_processing` but I haven't done this here. These would be an optimization beyond the issue at hand (improving block propagation times) and I think this PR is already doing enough. We can come back for that later. ## TODO - [x] Tidy, improve comments. - [x] ~~Try avoid computing proposer index in `per_block_processing`?~~
2021-02-15 07:17:52 +00:00
.ok_or(Error::SnapshotCacheLockTimeout)
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
.map(|mut snapshot_cache| {
Advance state to next slot after importing block (#2174) ## Issue Addressed NA ## Proposed Changes Add an optimization to perform `per_slot_processing` from the *leading-edge* of block processing to the *trailing-edge*. Ultimately, this allows us to import the block at slot `n` faster because we used the tail-end of slot `n - 1` to perform `per_slot_processing`. Additionally, add a "block proposer cache" which allows us to cache the block proposer for some epoch. Since we're now doing trailing-edge `per_slot_processing`, we can prime this cache with the values for the next epoch before those blocks arrive (assuming those blocks don't have some weird forking). There were several ancillary changes required to achieve this: - Remove the `state_root` field of `BeaconSnapshot`, since there's no need to know it on a `pre_state` and in all other cases we can just read it from `block.state_root()`. - This caused some "dust" changes of `snapshot.beacon_state_root` to `snapshot.beacon_state_root()`, where the `BeaconSnapshot::beacon_state_root()` func just reads the state root from the block. - Rename `types::ShuffingId` to `AttestationShufflingId`. I originally did this because I added a `ProposerShufflingId` struct which turned out to be not so useful. I thought this new name was more descriptive so I kept it. - Address https://github.com/ethereum/eth2.0-specs/pull/2196 - Add a debug log when we get a block with an unknown parent. There was previously no logging around this case. - Add a function to `BeaconState` to compute all proposers for an epoch without re-computing the active indices for each slot. ## Additional Info - ~~Blocked on #2173~~ - ~~Blocked on #2179~~ That PR was wrapped into this PR. - There's potentially some places where we could avoid computing the proposer indices in `per_block_processing` but I haven't done this here. These would be an optimization beyond the issue at hand (improving block propagation times) and I think this PR is already doing enough. We can come back for that later. ## TODO - [x] Tidy, improve comments. - [x] ~~Try avoid computing proposer index in `per_block_processing`?~~
2021-02-15 07:17:52 +00:00
snapshot_cache.insert(
BeaconSnapshot {
beacon_state: state,
beacon_block: signed_block,
beacon_block_root: block_root,
},
None,
&self.spec,
Advance state to next slot after importing block (#2174) ## Issue Addressed NA ## Proposed Changes Add an optimization to perform `per_slot_processing` from the *leading-edge* of block processing to the *trailing-edge*. Ultimately, this allows us to import the block at slot `n` faster because we used the tail-end of slot `n - 1` to perform `per_slot_processing`. Additionally, add a "block proposer cache" which allows us to cache the block proposer for some epoch. Since we're now doing trailing-edge `per_slot_processing`, we can prime this cache with the values for the next epoch before those blocks arrive (assuming those blocks don't have some weird forking). There were several ancillary changes required to achieve this: - Remove the `state_root` field of `BeaconSnapshot`, since there's no need to know it on a `pre_state` and in all other cases we can just read it from `block.state_root()`. - This caused some "dust" changes of `snapshot.beacon_state_root` to `snapshot.beacon_state_root()`, where the `BeaconSnapshot::beacon_state_root()` func just reads the state root from the block. - Rename `types::ShuffingId` to `AttestationShufflingId`. I originally did this because I added a `ProposerShufflingId` struct which turned out to be not so useful. I thought this new name was more descriptive so I kept it. - Address https://github.com/ethereum/eth2.0-specs/pull/2196 - Add a debug log when we get a block with an unknown parent. There was previously no logging around this case. - Add a function to `BeaconState` to compute all proposers for an epoch without re-computing the active indices for each slot. ## Additional Info - ~~Blocked on #2173~~ - ~~Blocked on #2179~~ That PR was wrapped into this PR. - There's potentially some places where we could avoid computing the proposer indices in `per_block_processing` but I haven't done this here. These would be an optimization beyond the issue at hand (improving block propagation times) and I think this PR is already doing enough. We can come back for that later. ## TODO - [x] Tidy, improve comments. - [x] ~~Try avoid computing proposer index in `per_block_processing`?~~
2021-02-15 07:17:52 +00:00
)
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
})
Advance state to next slot after importing block (#2174) ## Issue Addressed NA ## Proposed Changes Add an optimization to perform `per_slot_processing` from the *leading-edge* of block processing to the *trailing-edge*. Ultimately, this allows us to import the block at slot `n` faster because we used the tail-end of slot `n - 1` to perform `per_slot_processing`. Additionally, add a "block proposer cache" which allows us to cache the block proposer for some epoch. Since we're now doing trailing-edge `per_slot_processing`, we can prime this cache with the values for the next epoch before those blocks arrive (assuming those blocks don't have some weird forking). There were several ancillary changes required to achieve this: - Remove the `state_root` field of `BeaconSnapshot`, since there's no need to know it on a `pre_state` and in all other cases we can just read it from `block.state_root()`. - This caused some "dust" changes of `snapshot.beacon_state_root` to `snapshot.beacon_state_root()`, where the `BeaconSnapshot::beacon_state_root()` func just reads the state root from the block. - Rename `types::ShuffingId` to `AttestationShufflingId`. I originally did this because I added a `ProposerShufflingId` struct which turned out to be not so useful. I thought this new name was more descriptive so I kept it. - Address https://github.com/ethereum/eth2.0-specs/pull/2196 - Add a debug log when we get a block with an unknown parent. There was previously no logging around this case. - Add a function to `BeaconState` to compute all proposers for an epoch without re-computing the active indices for each slot. ## Additional Info - ~~Blocked on #2173~~ - ~~Blocked on #2179~~ That PR was wrapped into this PR. - There's potentially some places where we could avoid computing the proposer indices in `per_block_processing` but I haven't done this here. These would be an optimization beyond the issue at hand (improving block propagation times) and I think this PR is already doing enough. We can come back for that later. ## TODO - [x] Tidy, improve comments. - [x] ~~Try avoid computing proposer index in `per_block_processing`?~~
2021-02-15 07:17:52 +00:00
.unwrap_or_else(|e| {
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
error!(
self.log,
Advance state to next slot after importing block (#2174) ## Issue Addressed NA ## Proposed Changes Add an optimization to perform `per_slot_processing` from the *leading-edge* of block processing to the *trailing-edge*. Ultimately, this allows us to import the block at slot `n` faster because we used the tail-end of slot `n - 1` to perform `per_slot_processing`. Additionally, add a "block proposer cache" which allows us to cache the block proposer for some epoch. Since we're now doing trailing-edge `per_slot_processing`, we can prime this cache with the values for the next epoch before those blocks arrive (assuming those blocks don't have some weird forking). There were several ancillary changes required to achieve this: - Remove the `state_root` field of `BeaconSnapshot`, since there's no need to know it on a `pre_state` and in all other cases we can just read it from `block.state_root()`. - This caused some "dust" changes of `snapshot.beacon_state_root` to `snapshot.beacon_state_root()`, where the `BeaconSnapshot::beacon_state_root()` func just reads the state root from the block. - Rename `types::ShuffingId` to `AttestationShufflingId`. I originally did this because I added a `ProposerShufflingId` struct which turned out to be not so useful. I thought this new name was more descriptive so I kept it. - Address https://github.com/ethereum/eth2.0-specs/pull/2196 - Add a debug log when we get a block with an unknown parent. There was previously no logging around this case. - Add a function to `BeaconState` to compute all proposers for an epoch without re-computing the active indices for each slot. ## Additional Info - ~~Blocked on #2173~~ - ~~Blocked on #2179~~ That PR was wrapped into this PR. - There's potentially some places where we could avoid computing the proposer indices in `per_block_processing` but I haven't done this here. These would be an optimization beyond the issue at hand (improving block propagation times) and I think this PR is already doing enough. We can come back for that later. ## TODO - [x] Tidy, improve comments. - [x] ~~Try avoid computing proposer index in `per_block_processing`?~~
2021-02-15 07:17:52 +00:00
"Failed to insert snapshot";
"error" => ?e,
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
"task" => "process block"
);
});
self.head_tracker
.register_block(block_root, parent_root, slot);
Advance state to next slot after importing block (#2174) ## Issue Addressed NA ## Proposed Changes Add an optimization to perform `per_slot_processing` from the *leading-edge* of block processing to the *trailing-edge*. Ultimately, this allows us to import the block at slot `n` faster because we used the tail-end of slot `n - 1` to perform `per_slot_processing`. Additionally, add a "block proposer cache" which allows us to cache the block proposer for some epoch. Since we're now doing trailing-edge `per_slot_processing`, we can prime this cache with the values for the next epoch before those blocks arrive (assuming those blocks don't have some weird forking). There were several ancillary changes required to achieve this: - Remove the `state_root` field of `BeaconSnapshot`, since there's no need to know it on a `pre_state` and in all other cases we can just read it from `block.state_root()`. - This caused some "dust" changes of `snapshot.beacon_state_root` to `snapshot.beacon_state_root()`, where the `BeaconSnapshot::beacon_state_root()` func just reads the state root from the block. - Rename `types::ShuffingId` to `AttestationShufflingId`. I originally did this because I added a `ProposerShufflingId` struct which turned out to be not so useful. I thought this new name was more descriptive so I kept it. - Address https://github.com/ethereum/eth2.0-specs/pull/2196 - Add a debug log when we get a block with an unknown parent. There was previously no logging around this case. - Add a function to `BeaconState` to compute all proposers for an epoch without re-computing the active indices for each slot. ## Additional Info - ~~Blocked on #2173~~ - ~~Blocked on #2179~~ That PR was wrapped into this PR. - There's potentially some places where we could avoid computing the proposer indices in `per_block_processing` but I haven't done this here. These would be an optimization beyond the issue at hand (improving block propagation times) and I think this PR is already doing enough. We can come back for that later. ## TODO - [x] Tidy, improve comments. - [x] ~~Try avoid computing proposer index in `per_block_processing`?~~
2021-02-15 07:17:52 +00:00
// Send an event to the `events` endpoint after fully processing the block.
if let Some(event_handler) = self.event_handler.as_ref() {
if event_handler.has_block_subscribers() {
event_handler.register(EventKind::Block(SseBlock {
slot,
block: block_root,
Add `execution_optimistic` flag to HTTP responses (#3070) ## Issue Addressed #3031 ## Proposed Changes Updates the following API endpoints to conform with https://github.com/ethereum/beacon-APIs/pull/190 and https://github.com/ethereum/beacon-APIs/pull/196 - [x] `beacon/states/{state_id}/root` - [x] `beacon/states/{state_id}/fork` - [x] `beacon/states/{state_id}/finality_checkpoints` - [x] `beacon/states/{state_id}/validators` - [x] `beacon/states/{state_id}/validators/{validator_id}` - [x] `beacon/states/{state_id}/validator_balances` - [x] `beacon/states/{state_id}/committees` - [x] `beacon/states/{state_id}/sync_committees` - [x] `beacon/headers` - [x] `beacon/headers/{block_id}` - [x] `beacon/blocks/{block_id}` - [x] `beacon/blocks/{block_id}/root` - [x] `beacon/blocks/{block_id}/attestations` - [x] `debug/beacon/states/{state_id}` - [x] `debug/beacon/heads` - [x] `validator/duties/attester/{epoch}` - [x] `validator/duties/proposer/{epoch}` - [x] `validator/duties/sync/{epoch}` Updates the following Server-Sent Events: - [x] `events?topics=head` - [x] `events?topics=block` - [x] `events?topics=finalized_checkpoint` - [x] `events?topics=chain_reorg` ## Backwards Incompatible There is a very minor breaking change with the way the API now handles requests to `beacon/blocks/{block_id}/root` and `beacon/states/{state_id}/root` when `block_id` or `state_id` is the `Root` variant of `BlockId` and `StateId` respectively. Previously a request to a non-existent root would simply echo the root back to the requester: ``` curl "http://localhost:5052/eth/v1/beacon/states/0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa/root" {"data":{"root":"0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"}} ``` Now it will return a `404`: ``` curl "http://localhost:5052/eth/v1/beacon/blocks/0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa/root" {"code":404,"message":"NOT_FOUND: beacon block with root 0xaaaa…aaaa","stacktraces":[]} ``` In addition to this is the block root `0x0000000000000000000000000000000000000000000000000000000000000000` previously would return the genesis block. It will now return a `404`: ``` curl "http://localhost:5052/eth/v1/beacon/blocks/0x0000000000000000000000000000000000000000000000000000000000000000" {"code":404,"message":"NOT_FOUND: beacon block with root 0x0000…0000","stacktraces":[]} ``` ## Additional Info - `execution_optimistic` is always set, and will return `false` pre-Bellatrix. I am also open to the idea of doing something like `#[serde(skip_serializing_if = "Option::is_none")]`. - The value of `execution_optimistic` is set to `false` where possible. Any computation that is reliant on the `head` will simply use the `ExecutionStatus` of the head (unless the head block is pre-Bellatrix). Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-07-25 08:23:00 +00:00
execution_optimistic: payload_verification_status.is_optimistic(),
}));
}
}
Fix block processing blowup, upgrade metrics (#500) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Begin metrics refactor * Move beacon_chain to new metrics structure. * Make metrics not panic if already defined * Use global prometheus gather at rest api * Unify common metric fns into a crate * Add heavy metering to block processing * Remove hypen from prometheus metric name * Add more beacon chain metrics * Add beacon chain persistence metric * Prune op pool on finalization * Add extra prom beacon chain metrics * Prefix BeaconChain metrics with "beacon_" * Add more store metrics * Add basic metrics to libp2p * Add metrics to HTTP server * Remove old `http_server` crate * Update metrics names to be more like standard * Fix broken beacon chain metrics, add slot clock metrics * Add lighthouse_metrics gather fn * Remove http args * Fix wrong state given to op pool prune * Make prom metric names more consistent * Add more metrics, tidy existing metrics * Fix store block read metrics * Tidy attestation metrics * Fix minor PR comments * Allow travis failures on beta (see desc) There's a non-backward compatible change in `cargo fmt`. Stable and beta do not agree. * Tidy `lighthouse_metrics` docs * Fix typo
2019-08-19 11:02:34 +00:00
metrics::stop_timer(db_write_timer);
metrics::inc_counter(&metrics::BLOCK_PROCESSING_SUCCESSES);
let block_delay_total = get_slot_delay_ms(block_time_imported, slot, &self.slot_clock);
// Do not write to the cache for blocks older than 2 epochs, this helps reduce writes to
// the cache during sync.
if block_delay_total < self.slot_clock.slot_duration() * 64 {
// Store the timestamp of the block being imported into the cache.
self.block_times_cache.write().set_time_imported(
block_root,
current_slot,
block_time_imported,
);
}
// Do not store metrics if the block was > 4 slots old, this helps prevent noise during
// sync.
if block_delay_total < self.slot_clock.slot_duration() * 4 {
// Observe the delay between when we observed the block and when we imported it.
let block_delays = self.block_times_cache.read().get_block_delays(
block_root,
self.slot_clock
.start_of(current_slot)
.unwrap_or_else(|| Duration::from_secs(0)),
);
metrics::observe_duration(
&metrics::BEACON_BLOCK_IMPORTED_OBSERVED_DELAY_TIME,
block_delays
.imported
.unwrap_or_else(|| Duration::from_secs(0)),
);
}
Deposit Cache Finalization & Fast WS Sync (#2915) ## Summary The deposit cache now has the ability to finalize deposits. This will cause it to drop unneeded deposit logs and hashes in the deposit Merkle tree that are no longer required to construct deposit proofs. The cache is finalized whenever the latest finalized checkpoint has a new `Eth1Data` with all deposits imported. This has three benefits: 1. Improves the speed of constructing Merkle proofs for deposits as we can just replay deposits since the last finalized checkpoint instead of all historical deposits when re-constructing the Merkle tree. 2. Significantly faster weak subjectivity sync as the deposit cache can be transferred to the newly syncing node in compressed form. The Merkle tree that stores `N` finalized deposits requires a maximum of `log2(N)` hashes. The newly syncing node then only needs to download deposits since the last finalized checkpoint to have a full tree. 3. Future proofing in preparation for [EIP-4444](https://eips.ethereum.org/EIPS/eip-4444) as execution nodes will no longer be required to store logs permanently so we won't always have all historical logs available to us. ## More Details Image to illustrate how the deposit contract merkle tree evolves and finalizes along with the resulting `DepositTreeSnapshot` ![image](https://user-images.githubusercontent.com/37123614/151465302-5fc56284-8a69-4998-b20e-45db3934ac70.png) ## Other Considerations I've changed the structure of the `SszDepositCache` so once you load & save your database from this version of lighthouse, you will no longer be able to load it from older versions. Co-authored-by: ethDreamer <37123614+ethDreamer@users.noreply.github.com>
2022-10-30 04:04:24 +00:00
// Do not write to eth1 finalization cache for blocks older than 5 epochs
// this helps reduce noise during sync
if block_delay_total
< self.slot_clock.slot_duration() * 5 * (T::EthSpec::slots_per_epoch() as u32)
{
let parent_block_epoch = parent_block.slot().epoch(T::EthSpec::slots_per_epoch());
if parent_block_epoch < current_epoch {
// we've crossed epoch boundary, store Eth1FinalizationData
let (checkpoint, eth1_finalization_data) =
if current_slot % T::EthSpec::slots_per_epoch() == 0 {
// current block is the checkpoint
(
Checkpoint {
epoch: current_epoch,
root: block_root,
},
current_eth1_finalization_data,
)
} else {
// parent block is the checkpoint
(
Checkpoint {
epoch: current_epoch,
root: parent_block.canonical_root(),
},
parent_eth1_finalization_data,
)
};
if let Some(finalized_eth1_data) = self
.eth1_finalization_cache
.try_write_for(ETH1_FINALIZATION_CACHE_LOCK_TIMEOUT)
.and_then(|mut cache| {
cache.insert(checkpoint, eth1_finalization_data);
cache.finalize(&current_finalized_checkpoint)
})
{
if let Some(eth1_chain) = self.eth1_chain.as_ref() {
let finalized_deposit_count = finalized_eth1_data.deposit_count;
eth1_chain.finalize_eth1_data(finalized_eth1_data);
debug!(
self.log,
"called eth1_chain.finalize_eth1_data()";
"epoch" => current_finalized_checkpoint.epoch,
"deposit count" => finalized_deposit_count,
);
}
}
}
}
// Inform the unknown block cache, in case it was waiting on this block.
self.pre_finalization_block_cache
.block_processed(block_root);
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
Ok(block_root)
}
Run fork choice before block proposal (#3168) ## Issue Addressed Upcoming spec change https://github.com/ethereum/consensus-specs/pull/2878 ## Proposed Changes 1. Run fork choice at the start of every slot, and wait for this run to complete before proposing a block. 2. As an optimisation, also run fork choice 3/4 of the way through the slot (at 9s), _dequeueing attestations for the next slot_. 3. Remove the fork choice run from the state advance timer that occurred before advancing the state. ## Additional Info ### Block Proposal Accuracy This change makes us more likely to propose on top of the correct head in the presence of re-orgs with proposer boost in play. The main scenario that this change is designed to address is described in the linked spec issue. ### Attestation Accuracy This change _also_ makes us more likely to attest to the correct head. Currently in the case of a skipped slot at `slot` we only run fork choice 9s into `slot - 1`. This means the attestations from `slot - 1` aren't taken into consideration, and any boost applied to the block from `slot - 1` is not removed (it should be). In the language of the linked spec issue, this means we are liable to attest to C, even when the majority voting weight has already caused a re-org to B. ### Why remove the call before the state advance? If we've run fork choice at the start of the slot then it has already dequeued all the attestations from the previous slot, which are the only ones eligible to influence the head in the current slot. Running fork choice again is unnecessary (unless we run it for the next slot and try to pre-empt a re-org, but I don't currently think this is a great idea). ### Performance Based on Prater testing this adds about 5-25ms of runtime to block proposal times, which are 500-1000ms on average (and spike to 5s+ sometimes due to state handling issues :cry: ). I believe this is a small enough penalty to enable it by default, with the option to disable it via the new flag `--fork-choice-before-proposal-timeout 0`. Upcoming work on block packing and state representation will also reduce block production times in general, while removing the spikes. ### Implementation Fork choice gets invoked at the start of the slot via the `per_slot_task` function called from the slot timer. It then uses a condition variable to signal to block production that fork choice has been updated. This is a bit funky, but it seems to work. One downside of the timer-based approach is that it doesn't happen automatically in most of the tests. The test added by this PR has to trigger the run manually.
2022-05-20 05:02:11 +00:00
/// If configured, wait for the fork choice run at the start of the slot to complete.
fn wait_for_fork_choice_before_block_production(
self: &Arc<Self>,
slot: Slot,
) -> Result<(), BlockProductionError> {
if let Some(rx) = &self.fork_choice_signal_rx {
let current_slot = self
.slot()
.map_err(|_| BlockProductionError::UnableToReadSlot)?;
let timeout = Duration::from_millis(self.config.fork_choice_before_proposal_timeout_ms);
if slot == current_slot || slot == current_slot + 1 {
match rx.wait_for_fork_choice(slot, timeout) {
ForkChoiceWaitResult::Success(fc_slot) => {
debug!(
self.log,
"Fork choice successfully updated before block production";
"slot" => slot,
"fork_choice_slot" => fc_slot,
);
}
ForkChoiceWaitResult::Behind(fc_slot) => {
warn!(
self.log,
"Fork choice notifier out of sync with block production";
"fork_choice_slot" => fc_slot,
"slot" => slot,
"message" => "this block may be orphaned",
);
}
ForkChoiceWaitResult::TimeOut => {
warn!(
self.log,
"Timed out waiting for fork choice before proposal";
"message" => "this block may be orphaned",
);
}
}
} else {
error!(
self.log,
"Producing block at incorrect slot";
"block_slot" => slot,
"current_slot" => current_slot,
"message" => "check clock sync, this block may be orphaned",
);
}
}
Ok(())
}
/// Produce a new block at the given `slot`.
///
/// The produced block will not be inherently valid, it must be signed by a block producer.
/// Block signing is out of the scope of this function and should be done by a separate program.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub async fn produce_block<Payload: ExecPayload<T::EthSpec>>(
Run fork choice before block proposal (#3168) ## Issue Addressed Upcoming spec change https://github.com/ethereum/consensus-specs/pull/2878 ## Proposed Changes 1. Run fork choice at the start of every slot, and wait for this run to complete before proposing a block. 2. As an optimisation, also run fork choice 3/4 of the way through the slot (at 9s), _dequeueing attestations for the next slot_. 3. Remove the fork choice run from the state advance timer that occurred before advancing the state. ## Additional Info ### Block Proposal Accuracy This change makes us more likely to propose on top of the correct head in the presence of re-orgs with proposer boost in play. The main scenario that this change is designed to address is described in the linked spec issue. ### Attestation Accuracy This change _also_ makes us more likely to attest to the correct head. Currently in the case of a skipped slot at `slot` we only run fork choice 9s into `slot - 1`. This means the attestations from `slot - 1` aren't taken into consideration, and any boost applied to the block from `slot - 1` is not removed (it should be). In the language of the linked spec issue, this means we are liable to attest to C, even when the majority voting weight has already caused a re-org to B. ### Why remove the call before the state advance? If we've run fork choice at the start of the slot then it has already dequeued all the attestations from the previous slot, which are the only ones eligible to influence the head in the current slot. Running fork choice again is unnecessary (unless we run it for the next slot and try to pre-empt a re-org, but I don't currently think this is a great idea). ### Performance Based on Prater testing this adds about 5-25ms of runtime to block proposal times, which are 500-1000ms on average (and spike to 5s+ sometimes due to state handling issues :cry: ). I believe this is a small enough penalty to enable it by default, with the option to disable it via the new flag `--fork-choice-before-proposal-timeout 0`. Upcoming work on block packing and state representation will also reduce block production times in general, while removing the spikes. ### Implementation Fork choice gets invoked at the start of the slot via the `per_slot_task` function called from the slot timer. It then uses a condition variable to signal to block production that fork choice has been updated. This is a bit funky, but it seems to work. One downside of the timer-based approach is that it doesn't happen automatically in most of the tests. The test added by this PR has to trigger the run manually.
2022-05-20 05:02:11 +00:00
self: &Arc<Self>,
randao_reveal: Signature,
slot: Slot,
validator_graffiti: Option<Graffiti>,
) -> Result<BeaconBlockAndState<T::EthSpec, Payload>, BlockProductionError> {
self.produce_block_with_verification(
randao_reveal,
slot,
validator_graffiti,
ProduceBlockVerification::VerifyRandao,
)
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
.await
}
/// Same as `produce_block` but allowing for configuration of RANDAO-verification.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub async fn produce_block_with_verification<Payload: ExecPayload<T::EthSpec>>(
Run fork choice before block proposal (#3168) ## Issue Addressed Upcoming spec change https://github.com/ethereum/consensus-specs/pull/2878 ## Proposed Changes 1. Run fork choice at the start of every slot, and wait for this run to complete before proposing a block. 2. As an optimisation, also run fork choice 3/4 of the way through the slot (at 9s), _dequeueing attestations for the next slot_. 3. Remove the fork choice run from the state advance timer that occurred before advancing the state. ## Additional Info ### Block Proposal Accuracy This change makes us more likely to propose on top of the correct head in the presence of re-orgs with proposer boost in play. The main scenario that this change is designed to address is described in the linked spec issue. ### Attestation Accuracy This change _also_ makes us more likely to attest to the correct head. Currently in the case of a skipped slot at `slot` we only run fork choice 9s into `slot - 1`. This means the attestations from `slot - 1` aren't taken into consideration, and any boost applied to the block from `slot - 1` is not removed (it should be). In the language of the linked spec issue, this means we are liable to attest to C, even when the majority voting weight has already caused a re-org to B. ### Why remove the call before the state advance? If we've run fork choice at the start of the slot then it has already dequeued all the attestations from the previous slot, which are the only ones eligible to influence the head in the current slot. Running fork choice again is unnecessary (unless we run it for the next slot and try to pre-empt a re-org, but I don't currently think this is a great idea). ### Performance Based on Prater testing this adds about 5-25ms of runtime to block proposal times, which are 500-1000ms on average (and spike to 5s+ sometimes due to state handling issues :cry: ). I believe this is a small enough penalty to enable it by default, with the option to disable it via the new flag `--fork-choice-before-proposal-timeout 0`. Upcoming work on block packing and state representation will also reduce block production times in general, while removing the spikes. ### Implementation Fork choice gets invoked at the start of the slot via the `per_slot_task` function called from the slot timer. It then uses a condition variable to signal to block production that fork choice has been updated. This is a bit funky, but it seems to work. One downside of the timer-based approach is that it doesn't happen automatically in most of the tests. The test added by this PR has to trigger the run manually.
2022-05-20 05:02:11 +00:00
self: &Arc<Self>,
randao_reveal: Signature,
slot: Slot,
validator_graffiti: Option<Graffiti>,
verification: ProduceBlockVerification,
) -> Result<BeaconBlockAndState<T::EthSpec, Payload>, BlockProductionError> {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// Part 1/2 (blocking)
//
// Load the parent state from disk.
let chain = self.clone();
let (state, state_root_opt) = self
.task_executor
.spawn_blocking_handle(
move || chain.load_state_for_block_production::<Payload>(slot),
"produce_partial_beacon_block",
)
.ok_or(BlockProductionError::ShuttingDown)?
.await
.map_err(BlockProductionError::TokioJoin)??;
// Part 2/2 (async, with some blocking components)
//
// Produce the block upon the state
self.produce_block_on_state::<Payload>(
state,
state_root_opt,
slot,
randao_reveal,
validator_graffiti,
verification,
)
.await
}
/// Load a beacon state from the database for block production. This is a long-running process
/// that should not be performed in an `async` context.
fn load_state_for_block_production<Payload: ExecPayload<T::EthSpec>>(
self: &Arc<Self>,
slot: Slot,
) -> Result<(BeaconState<T::EthSpec>, Option<Hash256>), BlockProductionError> {
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
metrics::inc_counter(&metrics::BLOCK_PRODUCTION_REQUESTS);
let _complete_timer = metrics::start_timer(&metrics::BLOCK_PRODUCTION_TIMES);
Run fork choice before block proposal (#3168) ## Issue Addressed Upcoming spec change https://github.com/ethereum/consensus-specs/pull/2878 ## Proposed Changes 1. Run fork choice at the start of every slot, and wait for this run to complete before proposing a block. 2. As an optimisation, also run fork choice 3/4 of the way through the slot (at 9s), _dequeueing attestations for the next slot_. 3. Remove the fork choice run from the state advance timer that occurred before advancing the state. ## Additional Info ### Block Proposal Accuracy This change makes us more likely to propose on top of the correct head in the presence of re-orgs with proposer boost in play. The main scenario that this change is designed to address is described in the linked spec issue. ### Attestation Accuracy This change _also_ makes us more likely to attest to the correct head. Currently in the case of a skipped slot at `slot` we only run fork choice 9s into `slot - 1`. This means the attestations from `slot - 1` aren't taken into consideration, and any boost applied to the block from `slot - 1` is not removed (it should be). In the language of the linked spec issue, this means we are liable to attest to C, even when the majority voting weight has already caused a re-org to B. ### Why remove the call before the state advance? If we've run fork choice at the start of the slot then it has already dequeued all the attestations from the previous slot, which are the only ones eligible to influence the head in the current slot. Running fork choice again is unnecessary (unless we run it for the next slot and try to pre-empt a re-org, but I don't currently think this is a great idea). ### Performance Based on Prater testing this adds about 5-25ms of runtime to block proposal times, which are 500-1000ms on average (and spike to 5s+ sometimes due to state handling issues :cry: ). I believe this is a small enough penalty to enable it by default, with the option to disable it via the new flag `--fork-choice-before-proposal-timeout 0`. Upcoming work on block packing and state representation will also reduce block production times in general, while removing the spikes. ### Implementation Fork choice gets invoked at the start of the slot via the `per_slot_task` function called from the slot timer. It then uses a condition variable to signal to block production that fork choice has been updated. This is a bit funky, but it seems to work. One downside of the timer-based approach is that it doesn't happen automatically in most of the tests. The test added by this PR has to trigger the run manually.
2022-05-20 05:02:11 +00:00
let fork_choice_timer = metrics::start_timer(&metrics::BLOCK_PRODUCTION_FORK_CHOICE_TIMES);
self.wait_for_fork_choice_before_block_production(slot)?;
drop(fork_choice_timer);
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
// Producing a block requires the tree hash cache, so clone a full state corresponding to
// the head from the snapshot cache. Unfortunately we can't move the snapshot out of the
// cache (which would be fast), because we need to re-process the block after it has been
// signed. If we miss the cache or we're producing a block that conflicts with the head,
// fall back to getting the head from `slot - 1`.
let state_load_timer = metrics::start_timer(&metrics::BLOCK_PRODUCTION_STATE_LOAD_TIMES);
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// Atomically read some values from the head whilst avoiding holding cached head `Arc` any
// longer than necessary.
let (head_slot, head_block_root) = {
let head = self.canonical_head.cached_head();
(head.head_slot(), head.head_block_root())
};
let (state, state_root_opt) = if head_slot < slot {
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
// Normal case: proposing a block atop the current head. Use the snapshot cache.
if let Some(pre_state) = self
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
.snapshot_cache
.try_read_for(BLOCK_PROCESSING_CACHE_LOCK_TIMEOUT)
.and_then(|snapshot_cache| {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
snapshot_cache.get_state_for_block_production(head_block_root)
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
})
{
(pre_state.pre_state, pre_state.state_root)
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
} else {
warn!(
self.log,
"Block production cache miss";
"message" => "this block is more likely to be orphaned",
"slot" => slot,
);
let state = self
.state_at_slot(slot - 1, StateSkipConfig::WithStateRoots)
.map_err(|_| BlockProductionError::UnableToProduceAtSlot(slot))?;
(state, None)
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
}
} else {
warn!(
self.log,
"Producing block that conflicts with head";
"message" => "this block is more likely to be orphaned",
"slot" => slot,
);
let state = self
.state_at_slot(slot - 1, StateSkipConfig::WithStateRoots)
.map_err(|_| BlockProductionError::UnableToProduceAtSlot(slot))?;
(state, None)
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
};
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
drop(state_load_timer);
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
Ok((state, state_root_opt))
}
/// Produce a block for some `slot` upon the given `state`.
///
/// Typically the `self.produce_block()` function should be used, instead of calling this
/// function directly. This function is useful for purposefully creating forks or blocks at
/// non-current slots.
///
/// If required, the given state will be advanced to the given `produce_at_slot`, then a block
/// will be produced at that slot height.
///
/// The provided `state_root_opt` should only ever be set to `Some` if the contained value is
/// equal to the root of `state`. Providing this value will serve as an optimization to avoid
/// performing a tree hash in some scenarios.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub async fn produce_block_on_state<Payload: ExecPayload<T::EthSpec>>(
self: &Arc<Self>,
state: BeaconState<T::EthSpec>,
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
state_root_opt: Option<Hash256>,
produce_at_slot: Slot,
randao_reveal: Signature,
validator_graffiti: Option<Graffiti>,
verification: ProduceBlockVerification,
) -> Result<BeaconBlockAndState<T::EthSpec, Payload>, BlockProductionError> {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// Part 1/3 (blocking)
//
// Perform the state advance and block-packing functions.
let chain = self.clone();
let mut partial_beacon_block = self
.task_executor
.spawn_blocking_handle(
move || {
chain.produce_partial_beacon_block(
state,
state_root_opt,
produce_at_slot,
randao_reveal,
validator_graffiti,
)
},
"produce_partial_beacon_block",
)
.ok_or(BlockProductionError::ShuttingDown)?
.await
.map_err(BlockProductionError::TokioJoin)??;
// Part 2/3 (async)
//
// Wait for the execution layer to return an execution payload (if one is required).
let prepare_payload_handle = partial_beacon_block.prepare_payload_handle.take();
let execution_payload = if let Some(prepare_payload_handle) = prepare_payload_handle {
let execution_payload = prepare_payload_handle
.await
.map_err(BlockProductionError::TokioJoin)?
.ok_or(BlockProductionError::ShuttingDown)??;
Some(execution_payload)
} else {
None
};
// Part 3/3 (blocking)
//
// Perform the final steps of combining all the parts and computing the state root.
let chain = self.clone();
self.task_executor
.spawn_blocking_handle(
move || {
chain.complete_partial_beacon_block(
partial_beacon_block,
execution_payload,
verification,
)
},
"complete_partial_beacon_block",
)
.ok_or(BlockProductionError::ShuttingDown)?
.await
.map_err(BlockProductionError::TokioJoin)?
}
fn produce_partial_beacon_block<Payload: ExecPayload<T::EthSpec>>(
self: &Arc<Self>,
mut state: BeaconState<T::EthSpec>,
state_root_opt: Option<Hash256>,
produce_at_slot: Slot,
randao_reveal: Signature,
validator_graffiti: Option<Graffiti>,
) -> Result<PartialBeaconBlock<T::EthSpec, Payload>, BlockProductionError> {
Eth1 Integration (#542) * Refactor to cache Eth1Data * Fix merge conflicts and minor refactorings * Rename Eth1Cache to Eth1DataCache * Refactor events subscription * Add deposits module to interface with BeaconChain deposits * Remove utils * Rename to types.rs and add trait constraints to Eth1DataFetcher * Confirm to trait constraints. Make Web3DataFetcher cloneable * Make fetcher object member of deposit and eth1_data cache and other fixes * Fix update_cache function * Move fetch_eth1_data to impl block * Fix deposit tests * Create Eth1 object for interfacing with Beacon chain * Add `run` function for running update_cache and subscribe_deposit_logs tasks * Add logging * Run `cargo fmt` and make tests pass * Convert sync functions to async * Add timeouts to web3 functions * Return futures from cache functions * Add failed chaining of futures * Working cache updation * Clean up tests and `update_cache` function * Refactor `get_eth1_data` functions to work with future returning functions * Refactor eth1 `run` function to work with modified `update_cache` api * Minor changes * Add distance parameter to `update_cache` * Fix tests and other minor fixes * Working integration with cache and deposits * Add merkle_tree construction, proof generation and verification code * Add function to construct and fetch Deposits for BeaconNode * Add error handling * Import ssz * Add error handling to eth1 cache and fix minor errors * Run rustfmt * Fix minor bug * Rename Eth1Error and change to Result<T> * Change deposit fetching mechanism from notification based to poll based * Add deposits from eth1 chain in a given range every `x` blocks * Modify `run` function to accommodate changes * Minor fixes * Fix formatting * Initial commit. web3 api working. * Tidied up lib. Add function for fetching logs. * Refactor with `Eth1DataFetcher` trait * Add parsing for deposit contract logs and get_eth1_data function * Add `get_eth1_votes` function * Refactor to cache Eth1Data * Fix merge conflicts and minor refactorings * Rename Eth1Cache to Eth1DataCache * Refactor events subscription * Add deposits module to interface with BeaconChain deposits * Remove utils * Rename to types.rs and add trait constraints to Eth1DataFetcher * Confirm to trait constraints. Make Web3DataFetcher cloneable * Make fetcher object member of deposit and eth1_data cache and other fixes * Fix update_cache function * Move fetch_eth1_data to impl block * Fix deposit tests * Create Eth1 object for interfacing with Beacon chain * Add `run` function for running update_cache and subscribe_deposit_logs tasks * Add logging * Run `cargo fmt` and make tests pass * Convert sync functions to async * Add timeouts to web3 functions * Return futures from cache functions * Add failed chaining of futures * Working cache updation * Clean up tests and `update_cache` function * Refactor `get_eth1_data` functions to work with future returning functions * Refactor eth1 `run` function to work with modified `update_cache` api * Minor changes * Add distance parameter to `update_cache` * Fix tests and other minor fixes * Working integration with cache and deposits * Add merkle_tree construction, proof generation and verification code * Add function to construct and fetch Deposits for BeaconNode * Add error handling * Import ssz * Add error handling to eth1 cache and fix minor errors * Run rustfmt * Fix minor bug * Rename Eth1Error and change to Result<T> * Change deposit fetching mechanism from notification based to poll based * Add deposits from eth1 chain in a given range every `x` blocks * Modify `run` function to accommodate changes * Minor fixes * Fix formatting * Fix merge issue * Refactor with `Config` struct. Remote `ContractConfig` * Rename eth1_chain crate to eth1 * Rename files and read abi file using `fs::read` * Move eth1 to lib * Remove unnecessary mutability constraint * Add `Web3Backend` for returning actual eth1 data * Refactor `get_eth1_votes` to return a Result * Delete `eth1_chain` crate * Return `Result` from `get_deposits` * Fix range of deposits to return to beacon chain * Add `get_block_height_by_hash` trait function * Add naive method for getting `previous_eth1_distance` * Add eth1 config params to main config * Add instructions for setting up eth1 testing environment * Add build script to fetch deposit contract abi * Contract ABI is part of compiled binary * Fix minor bugs * Move docs to lib * Add timeout to config * Remove print statements * Change warn to error * Fix typos * Removed prints in test and get timeout value from config * Fixed error types * Added logging to web3_fetcher * Refactor for modified web3 api * Fix minor stuff * Add build script * Tidy, hide eth1 integration tests behind flag * Add http crate * Add first stages of eth1_test_rig * Fix deposits on test rig * Fix bug with deposit count method * Add block hash getter to http eth1 * Clean eth1 http crate and tests * Add script to start ganache * Adds deposit tree to eth1-http * Extend deposit tree tests * Tidy tests in eth1-http * Add more detail to get block request * Add block cache to eth1-http * Rename deposit tree to deposit cache * Add inital updating to eth1-http * Tidy updater * Fix compile bugs in tests * Adds an Eth1DataCache builder * Reorg eth1-http files * Add (failing) tests for eth1 updater * Rename files, fix bug in eth1-http * Ensure that ganache timestamps are increasing * Fix bugs with getting eth1data ancestors * Improve eth1 testing, fix bugs * Add truncate method to block cache * Add pruning to block cache update process * Add tests for block pruning * Allow for dropping an expired cache. * Add more comments * Add first compiling version of deposit updater * Add common fn for getting range of required blocks * Add passing deposit update test * Improve tests * Fix block pruning bug * Add tests for running two updates at once * Add updater services to eth1 * Add deposit collection to beacon chain * Add incomplete builder experiments * Add first working version of beacon chain builder * Update test harness to new beacon chain type * Rename builder file, tidy * Add first working client builder * Progress further on client builder * Update becaon node binary to use client builder * Ensure release tests compile * Remove old eth1 crate * Add first pass of new lighthouse binary * Fix websocket server startup * Remove old binary code from beacon_node crate * Add first working beacon node tests * Add genesis crate, new eth1 cache_2 * Add Serivce to Eth1Cache * Refactor with general eth1 improvements * Add passing genesis test * Tidy, add comments * Add more comments to eth1 service * Add further eth1 progress * Fix some bugs with genesis * Fix eth1 bugs, make eth1 linking more efficient * Shift logic in genesis service * Add more comments to genesis service * Add gzip, max request values, timeouts to http * Update testnet parameters to suit goerli testnet * Add ability to vary Fork, fix custom spec * Be more explicit about deposit fork version * Start adding beacon chain eth1 option * Add more flexibility to prod client * Further runtime refactoring * Allow for starting from store * Add bootstrapping to client config * Add remote_beacon_node crate * Update eth1 service for more configurability * Update eth1 tests to use less runtimes * Patch issues with tests using too many files * Move dummy eth1 backend flag * Ensure all tests pass * Add ganache-cli to Dockerfile * Use a special docker hub image for testing * Appease clippy * Move validator client into lighthouse binary * Allow starting with dummy eth1 backend * Improve logging * Fix dummy eth1 backend from cli * Add extra testnet command * Ensure consistent spec in beacon node * Update eth1 rig to work on goerli * Tidy lcli, start adding support for yaml config * Add incomplete YamlConfig struct * Remove efforts at YamlConfig * Add incomplete eth1 voting. Blocked on spec issues * Add (untested) first pass at eth1 vote algo * Add tests for winning vote * Add more tests for eth1 chain * Add more eth1 voting tests * Added more eth1 voting testing * Change test name * Add more tests to eth1 chain * Tidy eth1 generics, add more tests * Improve comments * Tidy beacon_node tests * Tidy, rename JsonRpc.. to Caching.. * Tidy voting logic * Tidy builder docs * Add comments, tidy eth1 * Add more comments to eth1 * Fix bug with winning_vote * Add doc comments to the `ClientBuilder` * Remove commented-out code * Improve `ClientBuilder` docs * Add comments to client config * Add decoding test for `ClientConfig` * Remove unused `DepositSet` struct * Tidy `block_cache` * Remove commented out lines * Remove unused code in `eth1` crate * Remove old validator binary `main.rs` * Tidy, fix tests compile error * Add initial tests for get_deposits * Remove dead code in eth1_test_rig * Update TestingDepositBuilder * Add testing for getting eth1 deposits * Fix duplicate rand dep * Remove dead code * Remove accidentally-added files * Fix comment in eth1_genesis_service * Add .gitignore for eth1_test_rig * Fix bug in eth1_genesis_service * Remove dead code from eth2_config * Fix tabs/spaces in root Cargo.toml * Tidy eth1 crate * Allow for re-use of eth1 service after genesis * Update docs for new CLI * Change README gif * Tidy eth1 http module * Tidy eth1 service * Tidy environment crate * Remove unused file * Tidy, add comments * Remove commented-out code * Address majority of Michael's comments * Address other PR comments * Add link to issue alongside TODO
2019-11-15 03:47:51 +00:00
let eth1_chain = self
.eth1_chain
.as_ref()
.ok_or(BlockProductionError::NoEth1ChainConnection)?;
Eth1 Integration (#542) * Refactor to cache Eth1Data * Fix merge conflicts and minor refactorings * Rename Eth1Cache to Eth1DataCache * Refactor events subscription * Add deposits module to interface with BeaconChain deposits * Remove utils * Rename to types.rs and add trait constraints to Eth1DataFetcher * Confirm to trait constraints. Make Web3DataFetcher cloneable * Make fetcher object member of deposit and eth1_data cache and other fixes * Fix update_cache function * Move fetch_eth1_data to impl block * Fix deposit tests * Create Eth1 object for interfacing with Beacon chain * Add `run` function for running update_cache and subscribe_deposit_logs tasks * Add logging * Run `cargo fmt` and make tests pass * Convert sync functions to async * Add timeouts to web3 functions * Return futures from cache functions * Add failed chaining of futures * Working cache updation * Clean up tests and `update_cache` function * Refactor `get_eth1_data` functions to work with future returning functions * Refactor eth1 `run` function to work with modified `update_cache` api * Minor changes * Add distance parameter to `update_cache` * Fix tests and other minor fixes * Working integration with cache and deposits * Add merkle_tree construction, proof generation and verification code * Add function to construct and fetch Deposits for BeaconNode * Add error handling * Import ssz * Add error handling to eth1 cache and fix minor errors * Run rustfmt * Fix minor bug * Rename Eth1Error and change to Result<T> * Change deposit fetching mechanism from notification based to poll based * Add deposits from eth1 chain in a given range every `x` blocks * Modify `run` function to accommodate changes * Minor fixes * Fix formatting * Initial commit. web3 api working. * Tidied up lib. Add function for fetching logs. * Refactor with `Eth1DataFetcher` trait * Add parsing for deposit contract logs and get_eth1_data function * Add `get_eth1_votes` function * Refactor to cache Eth1Data * Fix merge conflicts and minor refactorings * Rename Eth1Cache to Eth1DataCache * Refactor events subscription * Add deposits module to interface with BeaconChain deposits * Remove utils * Rename to types.rs and add trait constraints to Eth1DataFetcher * Confirm to trait constraints. Make Web3DataFetcher cloneable * Make fetcher object member of deposit and eth1_data cache and other fixes * Fix update_cache function * Move fetch_eth1_data to impl block * Fix deposit tests * Create Eth1 object for interfacing with Beacon chain * Add `run` function for running update_cache and subscribe_deposit_logs tasks * Add logging * Run `cargo fmt` and make tests pass * Convert sync functions to async * Add timeouts to web3 functions * Return futures from cache functions * Add failed chaining of futures * Working cache updation * Clean up tests and `update_cache` function * Refactor `get_eth1_data` functions to work with future returning functions * Refactor eth1 `run` function to work with modified `update_cache` api * Minor changes * Add distance parameter to `update_cache` * Fix tests and other minor fixes * Working integration with cache and deposits * Add merkle_tree construction, proof generation and verification code * Add function to construct and fetch Deposits for BeaconNode * Add error handling * Import ssz * Add error handling to eth1 cache and fix minor errors * Run rustfmt * Fix minor bug * Rename Eth1Error and change to Result<T> * Change deposit fetching mechanism from notification based to poll based * Add deposits from eth1 chain in a given range every `x` blocks * Modify `run` function to accommodate changes * Minor fixes * Fix formatting * Fix merge issue * Refactor with `Config` struct. Remote `ContractConfig` * Rename eth1_chain crate to eth1 * Rename files and read abi file using `fs::read` * Move eth1 to lib * Remove unnecessary mutability constraint * Add `Web3Backend` for returning actual eth1 data * Refactor `get_eth1_votes` to return a Result * Delete `eth1_chain` crate * Return `Result` from `get_deposits` * Fix range of deposits to return to beacon chain * Add `get_block_height_by_hash` trait function * Add naive method for getting `previous_eth1_distance` * Add eth1 config params to main config * Add instructions for setting up eth1 testing environment * Add build script to fetch deposit contract abi * Contract ABI is part of compiled binary * Fix minor bugs * Move docs to lib * Add timeout to config * Remove print statements * Change warn to error * Fix typos * Removed prints in test and get timeout value from config * Fixed error types * Added logging to web3_fetcher * Refactor for modified web3 api * Fix minor stuff * Add build script * Tidy, hide eth1 integration tests behind flag * Add http crate * Add first stages of eth1_test_rig * Fix deposits on test rig * Fix bug with deposit count method * Add block hash getter to http eth1 * Clean eth1 http crate and tests * Add script to start ganache * Adds deposit tree to eth1-http * Extend deposit tree tests * Tidy tests in eth1-http * Add more detail to get block request * Add block cache to eth1-http * Rename deposit tree to deposit cache * Add inital updating to eth1-http * Tidy updater * Fix compile bugs in tests * Adds an Eth1DataCache builder * Reorg eth1-http files * Add (failing) tests for eth1 updater * Rename files, fix bug in eth1-http * Ensure that ganache timestamps are increasing * Fix bugs with getting eth1data ancestors * Improve eth1 testing, fix bugs * Add truncate method to block cache * Add pruning to block cache update process * Add tests for block pruning * Allow for dropping an expired cache. * Add more comments * Add first compiling version of deposit updater * Add common fn for getting range of required blocks * Add passing deposit update test * Improve tests * Fix block pruning bug * Add tests for running two updates at once * Add updater services to eth1 * Add deposit collection to beacon chain * Add incomplete builder experiments * Add first working version of beacon chain builder * Update test harness to new beacon chain type * Rename builder file, tidy * Add first working client builder * Progress further on client builder * Update becaon node binary to use client builder * Ensure release tests compile * Remove old eth1 crate * Add first pass of new lighthouse binary * Fix websocket server startup * Remove old binary code from beacon_node crate * Add first working beacon node tests * Add genesis crate, new eth1 cache_2 * Add Serivce to Eth1Cache * Refactor with general eth1 improvements * Add passing genesis test * Tidy, add comments * Add more comments to eth1 service * Add further eth1 progress * Fix some bugs with genesis * Fix eth1 bugs, make eth1 linking more efficient * Shift logic in genesis service * Add more comments to genesis service * Add gzip, max request values, timeouts to http * Update testnet parameters to suit goerli testnet * Add ability to vary Fork, fix custom spec * Be more explicit about deposit fork version * Start adding beacon chain eth1 option * Add more flexibility to prod client * Further runtime refactoring * Allow for starting from store * Add bootstrapping to client config * Add remote_beacon_node crate * Update eth1 service for more configurability * Update eth1 tests to use less runtimes * Patch issues with tests using too many files * Move dummy eth1 backend flag * Ensure all tests pass * Add ganache-cli to Dockerfile * Use a special docker hub image for testing * Appease clippy * Move validator client into lighthouse binary * Allow starting with dummy eth1 backend * Improve logging * Fix dummy eth1 backend from cli * Add extra testnet command * Ensure consistent spec in beacon node * Update eth1 rig to work on goerli * Tidy lcli, start adding support for yaml config * Add incomplete YamlConfig struct * Remove efforts at YamlConfig * Add incomplete eth1 voting. Blocked on spec issues * Add (untested) first pass at eth1 vote algo * Add tests for winning vote * Add more tests for eth1 chain * Add more eth1 voting tests * Added more eth1 voting testing * Change test name * Add more tests to eth1 chain * Tidy eth1 generics, add more tests * Improve comments * Tidy beacon_node tests * Tidy, rename JsonRpc.. to Caching.. * Tidy voting logic * Tidy builder docs * Add comments, tidy eth1 * Add more comments to eth1 * Fix bug with winning_vote * Add doc comments to the `ClientBuilder` * Remove commented-out code * Improve `ClientBuilder` docs * Add comments to client config * Add decoding test for `ClientConfig` * Remove unused `DepositSet` struct * Tidy `block_cache` * Remove commented out lines * Remove unused code in `eth1` crate * Remove old validator binary `main.rs` * Tidy, fix tests compile error * Add initial tests for get_deposits * Remove dead code in eth1_test_rig * Update TestingDepositBuilder * Add testing for getting eth1 deposits * Fix duplicate rand dep * Remove dead code * Remove accidentally-added files * Fix comment in eth1_genesis_service * Add .gitignore for eth1_test_rig * Fix bug in eth1_genesis_service * Remove dead code from eth2_config * Fix tabs/spaces in root Cargo.toml * Tidy eth1 crate * Allow for re-use of eth1 service after genesis * Update docs for new CLI * Change README gif * Tidy eth1 http module * Tidy eth1 service * Tidy environment crate * Remove unused file * Tidy, add comments * Remove commented-out code * Address majority of Michael's comments * Address other PR comments * Add link to issue alongside TODO
2019-11-15 03:47:51 +00:00
// It is invalid to try to produce a block using a state from a future slot.
if state.slot() > produce_at_slot {
return Err(BlockProductionError::StateSlotTooHigh {
produce_at_slot,
state_slot: state.slot(),
});
}
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
let slot_timer = metrics::start_timer(&metrics::BLOCK_PRODUCTION_SLOT_PROCESS_TIMES);
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
// Ensure the state has performed a complete transition into the required slot.
complete_state_advance(&mut state, state_root_opt, produce_at_slot, &self.spec)?;
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
drop(slot_timer);
2019-03-21 20:11:04 +00:00
state.build_committee_cache(RelativeEpoch::Current, &self.spec)?;
let parent_root = if state.slot() > 0 {
2019-06-07 06:46:07 +00:00
*state
.get_block_root(state.slot() - 1)
2019-06-07 06:46:07 +00:00
.map_err(|_| BlockProductionError::UnableToGetBlockRootFromState)?
} else {
state.latest_block_header().canonical_root()
2019-06-07 06:46:07 +00:00
};
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let proposer_index = state.get_beacon_proposer_index(state.slot(), &self.spec)? as u64;
let pubkey = state
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
.validators()
.get(proposer_index as usize)
.map(|v| v.pubkey)
.ok_or(BlockProductionError::BeaconChain(
BeaconChainError::ValidatorIndexUnknown(proposer_index as usize),
))?;
let builder_params = BuilderParams {
pubkey,
slot: state.slot(),
chain_health: self
.is_healthy(&parent_root)
.map_err(BlockProductionError::BeaconChain)?,
};
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// If required, start the process of loading an execution payload from the EL early. This
// allows it to run concurrently with things like attestation packing.
let prepare_payload_handle = match &state {
BeaconState::Base(_) | BeaconState::Altair(_) => None,
BeaconState::Merge(_) => {
let prepare_payload_handle =
get_execution_payload(self.clone(), &state, proposer_index, builder_params)?;
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
Some(prepare_payload_handle)
}
};
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
let (mut proposer_slashings, mut attester_slashings, mut voluntary_exits) =
self.op_pool.get_slashings_and_exits(&state, &self.spec);
Prepare for public testnet (#628) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Update to use v0.9.2 version of deposit contract * Add efforts to automate eth1 testnet deployment * Fix lcli testnet deployer * Modify bn CLI to parse eth2_testnet_dir * Progress with account_manager deposit tools * Make account manager submit deposits * Add password option for submitting deposits * Allow custom deposit amount * Add long names to lcli clap * Add password option to lcli deploy command * Fix minor bugs whilst testing * Address Michael's comments * Add refund-deposit-contract to lcli * Use time instead of skip count for denying long skips * Improve logging for eth1 * Fix bug with validator services exiting on error * Drop the block cache after genesis * Modify eth1 testnet config * Improve eth1 logging * Make validator wait until genesis time * Fix bug in eth1 voting * Add more logging to eth1 voting * Handle errors in eth1 http module * Set SECONDS_PER_DAY to sensible minimum * Shorten delay before testnet start * Ensure eth1 block is produced without any votes * Improve eth1 logging * Fix broken tests in eth1 * Tidy code in rest_api * Fix failing test in deposit_contract * Make CLI args more consistent * Change validator/duties endpoint * Add time-based skip slot limiting * Add new error type missed in previous commit * Add log when waiting for genesis * Refactor beacon node CLI * Remove unused dep * Add lcli eth1-genesis command * Fix bug in master merge * Apply clippy lints to beacon node * Add support for YamlConfig in Eth2TestnetDir * Upgrade tesnet deposit contract version * Remove unnecessary logging and correct formatting * Add a hardcoded eth2 testnet config * Ensure http server flag works. Overwrite configs with flags. * Ensure boot nodes are loaded from testnet dir * Fix account manager CLI bugs * Fix bugs with beacon node cli * Allow testnet dir without boot nodes * Write genesis state as SSZ * Remove ---/n from the start of testnet_dir files * Set default libp2p address * Tidy account manager CLI, add logging * Add check to see if testnet dir exists * Apply reviewers suggestions * Add HeadTracker struct * Add fork choice persistence * Shorten slot time for simulator * Add the /beacon/heads API endpoint * Update hardcoded testnet * Add tests for BeaconChain persistence + fix bugs * Extend BeaconChain persistence testing * Ensure chain is finalized b4 persistence tests * Ensure boot_enr.yaml is include in binary * Refactor beacon_chain_sim * Move files about in beacon sim * Update beacon_chain_sim * Fix bug with deposit inclusion * Increase log in genesis service, fix todo * Tidy sim, fix broken rest_api tests * Fix more broken tests * Update testnet * Fix broken rest api test * Tidy account manager CLI * Use tempdir for account manager * Stop hardcoded testnet dir from creating dir * Rename Eth2TestnetDir to Eth2TestnetConfig * Change hardcoded -> hard_coded * Tidy account manager * Add log to account manager * Tidy, ensure head tracker is loaded from disk * Tidy beacon chain builder * Tidy eth1_chain * Adds log support for simulator * Revert "Adds log support for simulator" This reverts commit ec77c66a052350f551db145cf20f213823428dd3. * Adds log support for simulator * Tidy after self-review * Change default log level * Address Michael's delicious PR comments * Fix off-by-one in tests
2019-12-03 04:28:57 +00:00
let eth1_data = eth1_chain.eth1_data_for_block_production(&state, &self.spec)?;
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let deposits = eth1_chain.deposits_for_block_inclusion(&state, &eth1_data, &self.spec)?;
Prepare for public testnet (#628) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Update to use v0.9.2 version of deposit contract * Add efforts to automate eth1 testnet deployment * Fix lcli testnet deployer * Modify bn CLI to parse eth2_testnet_dir * Progress with account_manager deposit tools * Make account manager submit deposits * Add password option for submitting deposits * Allow custom deposit amount * Add long names to lcli clap * Add password option to lcli deploy command * Fix minor bugs whilst testing * Address Michael's comments * Add refund-deposit-contract to lcli * Use time instead of skip count for denying long skips * Improve logging for eth1 * Fix bug with validator services exiting on error * Drop the block cache after genesis * Modify eth1 testnet config * Improve eth1 logging * Make validator wait until genesis time * Fix bug in eth1 voting * Add more logging to eth1 voting * Handle errors in eth1 http module * Set SECONDS_PER_DAY to sensible minimum * Shorten delay before testnet start * Ensure eth1 block is produced without any votes * Improve eth1 logging * Fix broken tests in eth1 * Tidy code in rest_api * Fix failing test in deposit_contract * Make CLI args more consistent * Change validator/duties endpoint * Add time-based skip slot limiting * Add new error type missed in previous commit * Add log when waiting for genesis * Refactor beacon node CLI * Remove unused dep * Add lcli eth1-genesis command * Fix bug in master merge * Apply clippy lints to beacon node * Add support for YamlConfig in Eth2TestnetDir * Upgrade tesnet deposit contract version * Remove unnecessary logging and correct formatting * Add a hardcoded eth2 testnet config * Ensure http server flag works. Overwrite configs with flags. * Ensure boot nodes are loaded from testnet dir * Fix account manager CLI bugs * Fix bugs with beacon node cli * Allow testnet dir without boot nodes * Write genesis state as SSZ * Remove ---/n from the start of testnet_dir files * Set default libp2p address * Tidy account manager CLI, add logging * Add check to see if testnet dir exists * Apply reviewers suggestions * Add HeadTracker struct * Add fork choice persistence * Shorten slot time for simulator * Add the /beacon/heads API endpoint * Update hardcoded testnet * Add tests for BeaconChain persistence + fix bugs * Extend BeaconChain persistence testing * Ensure chain is finalized b4 persistence tests * Ensure boot_enr.yaml is include in binary * Refactor beacon_chain_sim * Move files about in beacon sim * Update beacon_chain_sim * Fix bug with deposit inclusion * Increase log in genesis service, fix todo * Tidy sim, fix broken rest_api tests * Fix more broken tests * Update testnet * Fix broken rest api test * Tidy account manager CLI * Use tempdir for account manager * Stop hardcoded testnet dir from creating dir * Rename Eth2TestnetDir to Eth2TestnetConfig * Change hardcoded -> hard_coded * Tidy account manager * Add log to account manager * Tidy, ensure head tracker is loaded from disk * Tidy beacon chain builder * Tidy eth1_chain * Adds log support for simulator * Revert "Adds log support for simulator" This reverts commit ec77c66a052350f551db145cf20f213823428dd3. * Adds log support for simulator * Tidy after self-review * Change default log level * Address Michael's delicious PR comments * Fix off-by-one in tests
2019-12-03 04:28:57 +00:00
// Iterate through the naive aggregation pool and ensure all the attestations from there
// are included in the operation pool.
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
let unagg_import_timer =
metrics::start_timer(&metrics::BLOCK_PRODUCTION_UNAGGREGATED_TIMES);
for attestation in self.naive_aggregation_pool.read().iter() {
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
let import = |attestation: &Attestation<T::EthSpec>| {
let attesting_indices = get_attesting_indices_from_state(&state, attestation)?;
self.op_pool
.insert_attestation(attestation.clone(), attesting_indices)
};
if let Err(e) = import(attestation) {
// Don't stop block production if there's an error, just create a log.
error!(
self.log,
"Attestation did not transfer to op pool";
"reason" => ?e
);
}
}
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
drop(unagg_import_timer);
// Override the beacon node's graffiti with graffiti from the validator, if present.
let graffiti = match validator_graffiti {
Some(graffiti) => graffiti,
None => self.graffiti,
};
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
let attestation_packing_timer =
metrics::start_timer(&metrics::BLOCK_PRODUCTION_ATTESTATION_TIMES);
let mut prev_filter_cache = HashMap::new();
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
let prev_attestation_filter = |att: &AttestationRef<T::EthSpec>| {
self.filter_op_pool_attestation(&mut prev_filter_cache, att, &state)
};
let mut curr_filter_cache = HashMap::new();
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
let curr_attestation_filter = |att: &AttestationRef<T::EthSpec>| {
self.filter_op_pool_attestation(&mut curr_filter_cache, att, &state)
};
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
let mut attestations = self
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
.op_pool
.get_attestations(
&state,
prev_attestation_filter,
curr_attestation_filter,
&self.spec,
)
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
.map_err(BlockProductionError::OpPoolError)?;
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
drop(attestation_packing_timer);
Refactor op pool for speed and correctness (#3312) ## Proposed Changes This PR has two aims: to speed up attestation packing in the op pool, and to fix bugs in the verification of attester slashings, proposer slashings and voluntary exits. The changes are bundled into a single database schema upgrade (v12). Attestation packing is sped up by removing several inefficiencies: - No more recalculation of `attesting_indices` during packing. - No (unnecessary) examination of the `ParticipationFlags`: a bitfield suffices. See `RewardCache`. - No re-checking of attestation validity during packing: the `AttestationMap` provides attestations which are "correct by construction" (I have checked this using Hydra). - No SSZ re-serialization for the clunky `AttestationId` type (it can be removed in a future release). So far the speed-up seems to be roughly 2-10x, from 500ms down to 50-100ms. Verification of attester slashings, proposer slashings and voluntary exits is fixed by: - Tracking the `ForkVersion`s that were used to verify each message inside the `SigVerifiedOp`. This allows us to quickly re-verify that they match the head state's opinion of what the `ForkVersion` should be at the epoch(s) relevant to the message. - Storing the `SigVerifiedOp` on disk rather than the raw operation. This allows us to continue track the fork versions after a reboot. This is mostly contained in this commit 52bb1840ae5c4356a8fc3a51e5df23ed65ed2c7f. ## Additional Info The schema upgrade uses the justified state to re-verify attestations and compute `attesting_indices` for them. It will drop any attestations that fail to verify, by the logic that attestations are most valuable in the few slots after they're observed, and are probably stale and useless by the time a node restarts. Exits and proposer slashings and similarly re-verified to obtain `SigVerifiedOp`s. This PR contains a runtime killswitch `--paranoid-block-proposal` which opts out of all the optimisations in favour of closely verifying every included message. Although I'm quite sure that the optimisations are correct this flag could be useful in the event of an unforeseen emergency. Finally, you might notice that the `RewardCache` appears quite useless in its current form because it is only updated on the hot-path immediately before proposal. My hope is that in future we can shift calls to `RewardCache::update` into the background, e.g. while performing the state advance. It is also forward-looking to `tree-states` compatibility, where iterating and indexing `state.{previous,current}_epoch_participation` is expensive and needs to be minimised.
2022-08-29 09:10:26 +00:00
// If paranoid mode is enabled re-check the signatures of every included message.
// This will be a lot slower but guards against bugs in block production and can be
// quickly rolled out without a release.
if self.config.paranoid_block_proposal {
attestations.retain(|att| {
verify_attestation_for_block_inclusion(
&state,
att,
VerifySignatures::True,
&self.spec,
)
.map_err(|e| {
warn!(
self.log,
"Attempted to include an invalid attestation";
"err" => ?e,
"block_slot" => state.slot(),
"attestation" => ?att
);
})
.is_ok()
});
proposer_slashings.retain(|slashing| {
slashing
.clone()
.validate(&state, &self.spec)
.map_err(|e| {
warn!(
self.log,
"Attempted to include an invalid proposer slashing";
"err" => ?e,
"block_slot" => state.slot(),
"slashing" => ?slashing
);
})
.is_ok()
});
attester_slashings.retain(|slashing| {
slashing
.clone()
.validate(&state, &self.spec)
.map_err(|e| {
warn!(
self.log,
"Attempted to include an invalid attester slashing";
"err" => ?e,
"block_slot" => state.slot(),
"slashing" => ?slashing
);
})
.is_ok()
});
voluntary_exits.retain(|exit| {
exit.clone()
.validate(&state, &self.spec)
.map_err(|e| {
warn!(
self.log,
"Attempted to include an invalid proposer slashing";
"err" => ?e,
"block_slot" => state.slot(),
"exit" => ?exit
);
})
.is_ok()
});
}
let slot = state.slot();
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let sync_aggregate = if matches!(&state, BeaconState::Base(_)) {
None
} else {
let sync_aggregate = self
.op_pool
.get_sync_aggregate(&state)
.map_err(BlockProductionError::OpPoolError)?
.unwrap_or_else(|| {
warn!(
self.log,
"Producing block with no sync contributions";
"slot" => state.slot(),
);
SyncAggregate::new()
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
});
Some(sync_aggregate)
};
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
Ok(PartialBeaconBlock {
state,
slot,
proposer_index,
parent_root,
randao_reveal,
eth1_data,
graffiti,
proposer_slashings,
attester_slashings,
attestations,
deposits,
voluntary_exits,
sync_aggregate,
prepare_payload_handle,
})
}
fn complete_partial_beacon_block<Payload: ExecPayload<T::EthSpec>>(
&self,
partial_beacon_block: PartialBeaconBlock<T::EthSpec, Payload>,
execution_payload: Option<Payload>,
verification: ProduceBlockVerification,
) -> Result<BeaconBlockAndState<T::EthSpec, Payload>, BlockProductionError> {
let PartialBeaconBlock {
mut state,
slot,
proposer_index,
parent_root,
randao_reveal,
eth1_data,
graffiti,
proposer_slashings,
attester_slashings,
attestations,
deposits,
voluntary_exits,
sync_aggregate,
// We don't need the prepare payload handle since the `execution_payload` is passed into
// this function. We can assume that the handle has already been consumed in order to
// produce said `execution_payload`.
prepare_payload_handle: _,
} = partial_beacon_block;
[Merge] Implement `execution_layer` (#2635) * Checkout serde_utils from rayonism * Make eth1::http functions pub * Add bones of execution_layer * Modify decoding * Expose Transaction, cargo fmt * Add executePayload * Add all minimal spec endpoints * Start adding json rpc wrapper * Finish custom JSON response handler * Switch to new rpc sending method * Add first test * Fix camelCase * Finish adding tests * Begin threading execution layer into BeaconChain * Fix clippy lints * Fix clippy lints * Thread execution layer into ClientBuilder * Add CLI flags * Add block processing methods to ExecutionLayer * Add block_on to execution_layer * Integrate execute_payload * Add extra_data field * Begin implementing payload handle * Send consensus valid/invalid messages * Fix minor type in task_executor * Call forkchoiceUpdated * Add search for TTD block * Thread TTD into execution layer * Allow producing block with execution payload * Add LRU cache for execution blocks * Remove duplicate 0x on ssz_types serialization * Add tests for block getter methods * Add basic block generator impl * Add is_valid_terminal_block to EL * Verify merge block in block_verification * Partially implement --terminal-block-hash-override * Add terminal_block_hash to ChainSpec * Remove Option from terminal_block_hash in EL * Revert merge changes to consensus/fork_choice * Remove commented-out code * Add bones for handling RPC methods on test server * Add first ExecutionLayer tests * Add testing for finding terminal block * Prevent infinite loops * Add insert_merge_block to block gen * Add block gen test for pos blocks * Start adding payloads to block gen * Fix clippy lints * Add execution payload to block gen * Add execute_payload to block_gen * Refactor block gen * Add all routes to mock server * Use Uint256 for base_fee_per_gas * Add working execution chain build * Remove unused var * Revert "Use Uint256 for base_fee_per_gas" This reverts commit 6c88f19ac45db834dd4dbf7a3c6e7242c1c0f735. * Fix base_fee_for_gas Uint256 * Update execute payload handle * Improve testing, fix bugs * Fix default fee-recipient * Fix fee-recipient address (again) * Add check for terminal block, add comments, tidy * Apply suggestions from code review Co-authored-by: realbigsean <seananderson33@GMAIL.com> * Fix is_none on handle Drop * Remove commented-out tests Co-authored-by: realbigsean <seananderson33@GMAIL.com>
2021-09-29 22:14:15 +00:00
let inner_block = match &state {
BeaconState::Base(_) => BeaconBlock::Base(BeaconBlockBase {
slot,
proposer_index,
Update to Spec v0.10 (#817) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Disable/delete two outdated tests * Bump eth1 default vote warn to error * Delete outdated eth1 test Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-02-10 23:19:36 +00:00
parent_root,
state_root: Hash256::zero(),
body: BeaconBlockBodyBase {
Update to Spec v0.10 (#817) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Disable/delete two outdated tests * Bump eth1 default vote warn to error * Delete outdated eth1 test Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-02-10 23:19:36 +00:00
randao_reveal,
eth1_data,
graffiti,
Update to Spec v0.10 (#817) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Disable/delete two outdated tests * Bump eth1 default vote warn to error * Delete outdated eth1 test Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-02-10 23:19:36 +00:00
proposer_slashings: proposer_slashings.into(),
attester_slashings: attester_slashings.into(),
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
attestations: attestations.into(),
deposits: deposits.into(),
voluntary_exits: voluntary_exits.into(),
_phantom: PhantomData,
Update to Spec v0.10 (#817) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Disable/delete two outdated tests * Bump eth1 default vote warn to error * Delete outdated eth1 test Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-02-10 23:19:36 +00:00
},
}),
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
BeaconState::Altair(_) => BeaconBlock::Altair(BeaconBlockAltair {
slot,
proposer_index,
parent_root,
state_root: Hash256::zero(),
body: BeaconBlockBodyAltair {
randao_reveal,
eth1_data,
graffiti,
proposer_slashings: proposer_slashings.into(),
attester_slashings: attester_slashings.into(),
attestations: attestations.into(),
deposits: deposits.into(),
voluntary_exits: voluntary_exits.into(),
sync_aggregate: sync_aggregate
.ok_or(BlockProductionError::MissingSyncAggregate)?,
_phantom: PhantomData,
},
}),
BeaconState::Merge(_) => BeaconBlock::Merge(BeaconBlockMerge {
slot,
proposer_index,
parent_root,
state_root: Hash256::zero(),
body: BeaconBlockBodyMerge {
randao_reveal,
eth1_data,
graffiti,
proposer_slashings: proposer_slashings.into(),
attester_slashings: attester_slashings.into(),
attestations: attestations.into(),
deposits: deposits.into(),
voluntary_exits: voluntary_exits.into(),
sync_aggregate: sync_aggregate
.ok_or(BlockProductionError::MissingSyncAggregate)?,
execution_payload: execution_payload
.ok_or(BlockProductionError::MissingExecutionPayload)?,
},
}),
};
let block = SignedBeaconBlock::from_block(
inner_block,
// The block is not signed here, that is the task of a validator client.
Signature::empty(),
);
let block_size = block.ssz_bytes_len();
debug!(
self.log,
"Produced block on state";
"block_size" => block_size,
);
metrics::observe(&metrics::BLOCK_SIZE, block_size as f64);
if block_size > self.config.max_network_size {
return Err(BlockProductionError::BlockTooLarge(block_size));
}
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
let process_timer = metrics::start_timer(&metrics::BLOCK_PRODUCTION_PROCESS_TIMES);
let signature_strategy = match verification {
ProduceBlockVerification::VerifyRandao => BlockSignatureStrategy::VerifyRandao,
ProduceBlockVerification::NoVerification => BlockSignatureStrategy::NoVerification,
};
// Use a context without block root or proposer index so that both are checked.
let mut ctxt = ConsensusContext::new(block.slot());
per_block_processing(
&mut state,
&block,
signature_strategy,
VerifyBlockRoot::True,
&mut ctxt,
&self.spec,
)?;
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
drop(process_timer);
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
let state_root_timer = metrics::start_timer(&metrics::BLOCK_PRODUCTION_STATE_ROOT_TIMES);
let state_root = state.update_tree_hash_cache()?;
Optimise tree hash caching for block production (#2106) ## Proposed Changes `@potuz` on the Eth R&D Discord observed that Lighthouse blocks on Pyrmont were always arriving at other nodes after at least 1 second. Part of this could be due to processing and slow propagation, but metrics also revealed that the Lighthouse nodes were usually taking 400-600ms to even just produce a block before broadcasting it. I tracked the slowness down to the lack of a pre-built tree hash cache (THC) on the states being used for block production. This was due to using the head state for block production, which lacks a THC in order to keep fork choice fast (cloning a THC takes at least 30ms for 100k validators). This PR modifies block production to clone a state from the snapshot cache rather than the head, which speeds things up by 200-400ms by avoiding the tree hash cache rebuild. In practice this seems to have cut block production time down to 300ms or less. Ideally we could _remove_ the snapshot from the cache (and save the 30ms), but it is required for when we re-process the block after signing it with the validator client. ## Alternatives I experimented with 2 alternatives to this approach, before deciding on it: * Alternative 1: ensure the `head` has a tree hash cache. This is too slow, as it imposes a +30ms hit on fork choice, which currently takes ~5ms (with occasional spikes). * Alternative 2: use `Arc<BeaconSnapshot>` in the snapshot cache and share snapshots between the cache and the `head`. This made fork choice blazing fast (1ms), and block production the same as in this PR, but had a negative impact on block processing which I don't think is worth it. It ended up being necessary to clone the full state from the snapshot cache during block production, imposing the +30ms penalty there _as well_ as in block production. In contract, the approach in this PR should only impact block production, and it improves it! Yay for pareto improvements :tada: ## Additional Info This commit (ac59dfa) is currently running on all the Lighthouse Pyrmont nodes, and I've added a dashboard to the Pyrmont grafana instance with the metrics. In future work we should optimise the attestation packing, which consumes around 30-60ms and is now a substantial contributor to the total.
2020-12-21 06:29:39 +00:00
drop(state_root_timer);
let (mut block, _) = block.deconstruct();
*block.state_root_mut() = state_root;
Fix block processing blowup, upgrade metrics (#500) * Renamed fork_choice::process_attestation_from_block * Processing attestation in fork choice * Retrieving state from store and checking signature * Looser check on beacon state validity. * Cleaned up get_attestation_state * Expanded fork choice api to provide latest validator message. * Checking if the an attestation contains a latest message * Correct process_attestation error handling. * Copy paste error in comment fixed. * Tidy ancestor iterators * Getting attestation slot via helper method * Refactored attestation creation in test utils * Revert "Refactored attestation creation in test utils" This reverts commit 4d277fe4239a7194758b18fb5c00dfe0b8231306. * Integration tests for free attestation processing * Implicit conflicts resolved. * formatting * Do first pass on Grants code * Add another attestation processing test * Tidy attestation processing * Remove old code fragment * Add non-compiling half finished changes * Simplify, fix bugs, add tests for chain iters * Remove attestation processing from op pool * Fix bug with fork choice, tidy * Fix overly restrictive check in fork choice. * Ensure committee cache is build during attn proc * Ignore unknown blocks at fork choice * Various minor fixes * Make fork choice write lock in to read lock * Remove unused method * Tidy comments * Fix attestation prod. target roots change * Fix compile error in store iters * Reject any attestation prior to finalization * Begin metrics refactor * Move beacon_chain to new metrics structure. * Make metrics not panic if already defined * Use global prometheus gather at rest api * Unify common metric fns into a crate * Add heavy metering to block processing * Remove hypen from prometheus metric name * Add more beacon chain metrics * Add beacon chain persistence metric * Prune op pool on finalization * Add extra prom beacon chain metrics * Prefix BeaconChain metrics with "beacon_" * Add more store metrics * Add basic metrics to libp2p * Add metrics to HTTP server * Remove old `http_server` crate * Update metrics names to be more like standard * Fix broken beacon chain metrics, add slot clock metrics * Add lighthouse_metrics gather fn * Remove http args * Fix wrong state given to op pool prune * Make prom metric names more consistent * Add more metrics, tidy existing metrics * Fix store block read metrics * Tidy attestation metrics * Fix minor PR comments * Allow travis failures on beta (see desc) There's a non-backward compatible change in `cargo fmt`. Stable and beta do not agree. * Tidy `lighthouse_metrics` docs * Fix typo
2019-08-19 11:02:34 +00:00
metrics::inc_counter(&metrics::BLOCK_PRODUCTION_SUCCESSES);
trace!(
self.log,
"Produced beacon block";
"parent" => ?block.parent_root(),
"attestations" => block.body().attestations().len(),
"slot" => block.slot()
);
Ok((block, state))
}
/// This method must be called whenever an execution engine indicates that a payload is
/// invalid.
///
Rename random to prev_randao (#3040) ## Issue Addressed As discussed on last-night's consensus call, the testnets next week will target the [Kiln Spec v2](https://hackmd.io/@n0ble/kiln-spec). Presently, we support Kiln V1. V2 is backwards compatible, except for renaming `random` to `prev_randao` in: - https://github.com/ethereum/execution-apis/pull/180 - https://github.com/ethereum/consensus-specs/pull/2835 With this PR we'll no longer be compatible with the existing Kintsugi and Kiln testnets, however we'll be ready for the testnets next week. I raised this breaking change in the call last night, we are all keen to move forward and break things. We now target the [`merge-kiln-v2`](https://github.com/MariusVanDerWijden/go-ethereum/tree/merge-kiln-v2) branch for interop with Geth. This required adding the `--http.aauthport` to the tester to avoid a port conflict at startup. ### Changes to exec integration tests There's some change in the `merge-kiln-v2` version of Geth that means it can't compile on a vanilla Github runner. Bumping the `go` version on the runner solved this issue. Whilst addressing this, I refactored the `testing/execution_integration` crate to be a *binary* rather than a *library* with tests. This means that we don't need to run the `build.rs` and build Geth whenever someone runs `make lint` or `make test-release`. This is nice for everyday users, but it's also nice for CI so that we can have a specific runner for these tests and we don't need to ensure *all* runners support everything required to build all execution clients. ## More Info - [x] ~~EF tests are failing since the rename has broken some tests that reference the old field name. I have been told there will be new tests released in the coming days (25/02/22 or 26/02/22).~~
2022-03-03 02:10:57 +00:00
/// Fork choice will be run after the invalidation. The client may be shut down if the `op`
/// results in the justified checkpoint being invalidated.
///
Rename random to prev_randao (#3040) ## Issue Addressed As discussed on last-night's consensus call, the testnets next week will target the [Kiln Spec v2](https://hackmd.io/@n0ble/kiln-spec). Presently, we support Kiln V1. V2 is backwards compatible, except for renaming `random` to `prev_randao` in: - https://github.com/ethereum/execution-apis/pull/180 - https://github.com/ethereum/consensus-specs/pull/2835 With this PR we'll no longer be compatible with the existing Kintsugi and Kiln testnets, however we'll be ready for the testnets next week. I raised this breaking change in the call last night, we are all keen to move forward and break things. We now target the [`merge-kiln-v2`](https://github.com/MariusVanDerWijden/go-ethereum/tree/merge-kiln-v2) branch for interop with Geth. This required adding the `--http.aauthport` to the tester to avoid a port conflict at startup. ### Changes to exec integration tests There's some change in the `merge-kiln-v2` version of Geth that means it can't compile on a vanilla Github runner. Bumping the `go` version on the runner solved this issue. Whilst addressing this, I refactored the `testing/execution_integration` crate to be a *binary* rather than a *library* with tests. This means that we don't need to run the `build.rs` and build Geth whenever someone runs `make lint` or `make test-release`. This is nice for everyday users, but it's also nice for CI so that we can have a specific runner for these tests and we don't need to ensure *all* runners support everything required to build all execution clients. ## More Info - [x] ~~EF tests are failing since the rename has broken some tests that reference the old field name. I have been told there will be new tests released in the coming days (25/02/22 or 26/02/22).~~
2022-03-03 02:10:57 +00:00
/// See the documentation of `InvalidationOperation` for information about defining `op`.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub async fn process_invalid_execution_payload(
self: &Arc<Self>,
Rename random to prev_randao (#3040) ## Issue Addressed As discussed on last-night's consensus call, the testnets next week will target the [Kiln Spec v2](https://hackmd.io/@n0ble/kiln-spec). Presently, we support Kiln V1. V2 is backwards compatible, except for renaming `random` to `prev_randao` in: - https://github.com/ethereum/execution-apis/pull/180 - https://github.com/ethereum/consensus-specs/pull/2835 With this PR we'll no longer be compatible with the existing Kintsugi and Kiln testnets, however we'll be ready for the testnets next week. I raised this breaking change in the call last night, we are all keen to move forward and break things. We now target the [`merge-kiln-v2`](https://github.com/MariusVanDerWijden/go-ethereum/tree/merge-kiln-v2) branch for interop with Geth. This required adding the `--http.aauthport` to the tester to avoid a port conflict at startup. ### Changes to exec integration tests There's some change in the `merge-kiln-v2` version of Geth that means it can't compile on a vanilla Github runner. Bumping the `go` version on the runner solved this issue. Whilst addressing this, I refactored the `testing/execution_integration` crate to be a *binary* rather than a *library* with tests. This means that we don't need to run the `build.rs` and build Geth whenever someone runs `make lint` or `make test-release`. This is nice for everyday users, but it's also nice for CI so that we can have a specific runner for these tests and we don't need to ensure *all* runners support everything required to build all execution clients. ## More Info - [x] ~~EF tests are failing since the rename has broken some tests that reference the old field name. I have been told there will be new tests released in the coming days (25/02/22 or 26/02/22).~~
2022-03-03 02:10:57 +00:00
op: &InvalidationOperation,
) -> Result<(), Error> {
debug!(
self.log,
"Processing payload invalidation";
"op" => ?op,
);
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// Update the execution status in fork choice.
//
// Use a blocking task since it interacts with the `canonical_head` lock. Lock contention
// on the core executor is bad.
let chain = self.clone();
let inner_op = op.clone();
let fork_choice_result = self
.spawn_blocking_handle(
move || {
chain
.canonical_head
.fork_choice_write_lock()
.on_invalid_execution_payload(&inner_op)
},
"invalid_payload_fork_choice_update",
)
.await?;
// Update fork choice.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
if let Err(e) = fork_choice_result {
crit!(
self.log,
"Failed to process invalid payload";
"error" => ?e,
Rename random to prev_randao (#3040) ## Issue Addressed As discussed on last-night's consensus call, the testnets next week will target the [Kiln Spec v2](https://hackmd.io/@n0ble/kiln-spec). Presently, we support Kiln V1. V2 is backwards compatible, except for renaming `random` to `prev_randao` in: - https://github.com/ethereum/execution-apis/pull/180 - https://github.com/ethereum/consensus-specs/pull/2835 With this PR we'll no longer be compatible with the existing Kintsugi and Kiln testnets, however we'll be ready for the testnets next week. I raised this breaking change in the call last night, we are all keen to move forward and break things. We now target the [`merge-kiln-v2`](https://github.com/MariusVanDerWijden/go-ethereum/tree/merge-kiln-v2) branch for interop with Geth. This required adding the `--http.aauthport` to the tester to avoid a port conflict at startup. ### Changes to exec integration tests There's some change in the `merge-kiln-v2` version of Geth that means it can't compile on a vanilla Github runner. Bumping the `go` version on the runner solved this issue. Whilst addressing this, I refactored the `testing/execution_integration` crate to be a *binary* rather than a *library* with tests. This means that we don't need to run the `build.rs` and build Geth whenever someone runs `make lint` or `make test-release`. This is nice for everyday users, but it's also nice for CI so that we can have a specific runner for these tests and we don't need to ensure *all* runners support everything required to build all execution clients. ## More Info - [x] ~~EF tests are failing since the rename has broken some tests that reference the old field name. I have been told there will be new tests released in the coming days (25/02/22 or 26/02/22).~~
2022-03-03 02:10:57 +00:00
"latest_valid_ancestor" => ?op.latest_valid_ancestor(),
"block_root" => ?op.block_root(),
);
}
// Run fork choice since it's possible that the payload invalidation might result in a new
// head.
self.recompute_head_at_current_slot().await;
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// Obtain the justified root from fork choice.
//
// Use a blocking task since it interacts with the `canonical_head` lock. Lock contention
// on the core executor is bad.
let chain = self.clone();
let justified_block = self
.spawn_blocking_handle(
move || {
chain
.canonical_head
.fork_choice_read_lock()
.get_justified_block()
},
"invalid_payload_fork_choice_get_justified",
)
.await??;
if justified_block.execution_status.is_invalid() {
crit!(
self.log,
"The justified checkpoint is invalid";
"msg" => "ensure you are not connected to a malicious network. This error is not \
recoverable, please reach out to the lighthouse developers for assistance."
);
let mut shutdown_sender = self.shutdown_sender();
if let Err(e) = shutdown_sender.try_send(ShutdownReason::Failure(
INVALID_JUSTIFIED_PAYLOAD_SHUTDOWN_REASON,
)) {
crit!(
self.log,
"Unable to trigger client shut down";
"msg" => "shut down may already be under way",
"error" => ?e
);
}
// Return an error here to try and prevent progression by upstream functions.
return Err(Error::JustifiedPayloadInvalid {
justified_root: justified_block.root,
execution_block_hash: justified_block.execution_status.block_hash(),
});
}
Ok(())
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub fn block_is_known_to_fork_choice(&self, root: &Hash256) -> bool {
self.canonical_head
.fork_choice_read_lock()
.contains_block(root)
}
/// Determines the beacon proposer for the next slot. If that proposer is registered in the
/// `execution_layer`, provide the `execution_layer` with the necessary information to produce
/// `PayloadAttributes` for future calls to fork choice.
///
/// The `PayloadAttributes` are used by the EL to give it a look-ahead for preparing an optimal
/// set of transactions for a new `ExecutionPayload`.
///
/// This function will result in a call to `forkchoiceUpdated` on the EL if:
///
/// 1. We're in the tail-end of the slot (as defined by PAYLOAD_PREPARATION_LOOKAHEAD_FACTOR)
/// 2. The head block is one slot (or less) behind the prepare slot (e.g., we're preparing for
/// the next slot and the block at the current slot is already known).
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub async fn prepare_beacon_proposer(
self: &Arc<Self>,
current_slot: Slot,
) -> Result<(), Error> {
let prepare_slot = current_slot + 1;
let prepare_epoch = prepare_slot.epoch(T::EthSpec::slots_per_epoch());
// There's no need to run the proposer preparation routine before the bellatrix fork.
if self.slot_is_prior_to_bellatrix(prepare_slot) {
return Ok(());
}
let execution_layer = self
.execution_layer
.clone()
.ok_or(Error::ExecutionLayerMissing)?;
// Nothing to do if there are no proposers registered with the EL, exit early to avoid
// wasting cycles.
if !execution_layer.has_any_proposer_preparation_data().await {
return Ok(());
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// Atomically read some values from the canonical head, whilst avoiding holding the cached
// head `Arc` any longer than necessary.
//
// Use a blocking task since blocking the core executor on the canonical head read lock can
// block the core tokio executor.
let chain = self.clone();
let (head_slot, head_root, head_decision_root, head_random, forkchoice_update_params) =
self.spawn_blocking_handle(
move || {
let cached_head = chain.canonical_head.cached_head();
let head_block_root = cached_head.head_block_root();
let decision_root = cached_head
.snapshot
.beacon_state
.proposer_shuffling_decision_root(head_block_root)?;
Ok::<_, Error>((
cached_head.head_slot(),
head_block_root,
decision_root,
cached_head.head_random()?,
cached_head.forkchoice_update_parameters(),
))
},
"prepare_beacon_proposer_fork_choice_read",
)
.await??;
let head_epoch = head_slot.epoch(T::EthSpec::slots_per_epoch());
// Don't bother with proposer prep if the head is more than
// `PREPARE_PROPOSER_HISTORIC_EPOCHS` prior to the current slot.
//
// This prevents the routine from running during sync.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
if head_slot + T::EthSpec::slots_per_epoch() * PREPARE_PROPOSER_HISTORIC_EPOCHS
< current_slot
{
debug!(
self.log,
"Head too old for proposer prep";
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
"head_slot" => head_slot,
"current_slot" => current_slot,
);
return Ok(());
}
// Ensure that the shuffling decision root is correct relative to the epoch we wish to
// query.
let shuffling_decision_root = if head_epoch == prepare_epoch {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
head_decision_root
} else {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
head_root
};
// Read the proposer from the proposer cache.
let cached_proposer = self
.beacon_proposer_cache
.lock()
.get_slot::<T::EthSpec>(shuffling_decision_root, prepare_slot);
let proposer = if let Some(proposer) = cached_proposer {
proposer.index
} else {
if head_epoch + 2 < prepare_epoch {
warn!(
self.log,
"Skipping proposer preparation";
"msg" => "this is a non-critical issue that can happen on unhealthy nodes or \
networks.",
"prepare_epoch" => prepare_epoch,
"head_epoch" => head_epoch,
);
// Don't skip the head forward more than two epochs. This avoids burdening an
// unhealthy node.
//
// Although this node might miss out on preparing for a proposal, they should still
// be able to propose. This will prioritise beacon chain health over efficient
// packing of execution blocks.
return Ok(());
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let (proposers, decision_root, _, fork) =
compute_proposer_duties_from_head(prepare_epoch, self)?;
let proposer_index = prepare_slot.as_usize() % (T::EthSpec::slots_per_epoch() as usize);
let proposer = *proposers
.get(proposer_index)
.ok_or(BeaconChainError::NoProposerForSlot(prepare_slot))?;
self.beacon_proposer_cache.lock().insert(
prepare_epoch,
decision_root,
proposers,
fork,
)?;
// It's possible that the head changes whilst computing these duties. If so, abandon
// this routine since the change of head would have also spawned another instance of
// this routine.
//
// Exit now, after updating the cache.
if decision_root != shuffling_decision_root {
warn!(
self.log,
"Head changed during proposer preparation";
);
return Ok(());
}
proposer
};
// If the execution layer doesn't have any proposer data for this validator then we assume
// it's not connected to this BN and no action is required.
if !execution_layer
.has_proposer_preparation_data(proposer as u64)
.await
{
return Ok(());
}
let payload_attributes = PayloadAttributes {
timestamp: self
.slot_clock
.start_of(prepare_slot)
.ok_or(Error::InvalidSlot(prepare_slot))?
.as_secs(),
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
prev_randao: head_random,
suggested_fee_recipient: execution_layer
.get_suggested_fee_recipient(proposer as u64)
.await,
};
debug!(
self.log,
"Preparing beacon proposer";
"payload_attributes" => ?payload_attributes,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
"head_root" => ?head_root,
"prepare_slot" => prepare_slot,
"validator" => proposer,
);
let already_known = execution_layer
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
.insert_proposer(prepare_slot, head_root, proposer as u64, payload_attributes)
.await;
// Only push a log to the user if this is the first time we've seen this proposer for this
// slot.
if !already_known {
info!(
self.log,
"Prepared beacon proposer";
"already_known" => already_known,
"prepare_slot" => prepare_slot,
"validator" => proposer,
);
}
let till_prepare_slot =
if let Some(duration) = self.slot_clock.duration_to_slot(prepare_slot) {
duration
} else {
// `SlotClock::duration_to_slot` will return `None` when we are past the start
// of `prepare_slot`. Don't bother sending a `forkchoiceUpdated` in that case,
// it's too late.
//
// This scenario might occur on an overloaded/under-resourced node.
warn!(
self.log,
"Delayed proposer preparation";
"prepare_slot" => prepare_slot,
"validator" => proposer,
);
return Ok(());
};
// If either of the following are true, send a fork-choice update message to the
// EL:
//
// 1. We're in the tail-end of the slot (as defined by
// PAYLOAD_PREPARATION_LOOKAHEAD_FACTOR)
// 2. The head block is one slot (or less) behind the prepare slot (e.g., we're
// preparing for the next slot and the block at the current slot is already
// known).
if till_prepare_slot
<= self.slot_clock.slot_duration() / PAYLOAD_PREPARATION_LOOKAHEAD_FACTOR
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
|| head_slot + 1 >= prepare_slot
{
debug!(
self.log,
"Pushing update to prepare proposer";
"till_prepare_slot" => ?till_prepare_slot,
"prepare_slot" => prepare_slot
);
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
self.update_execution_engine_forkchoice(current_slot, forkchoice_update_params)
.await?;
[Merge] Implement `execution_layer` (#2635) * Checkout serde_utils from rayonism * Make eth1::http functions pub * Add bones of execution_layer * Modify decoding * Expose Transaction, cargo fmt * Add executePayload * Add all minimal spec endpoints * Start adding json rpc wrapper * Finish custom JSON response handler * Switch to new rpc sending method * Add first test * Fix camelCase * Finish adding tests * Begin threading execution layer into BeaconChain * Fix clippy lints * Fix clippy lints * Thread execution layer into ClientBuilder * Add CLI flags * Add block processing methods to ExecutionLayer * Add block_on to execution_layer * Integrate execute_payload * Add extra_data field * Begin implementing payload handle * Send consensus valid/invalid messages * Fix minor type in task_executor * Call forkchoiceUpdated * Add search for TTD block * Thread TTD into execution layer * Allow producing block with execution payload * Add LRU cache for execution blocks * Remove duplicate 0x on ssz_types serialization * Add tests for block getter methods * Add basic block generator impl * Add is_valid_terminal_block to EL * Verify merge block in block_verification * Partially implement --terminal-block-hash-override * Add terminal_block_hash to ChainSpec * Remove Option from terminal_block_hash in EL * Revert merge changes to consensus/fork_choice * Remove commented-out code * Add bones for handling RPC methods on test server * Add first ExecutionLayer tests * Add testing for finding terminal block * Prevent infinite loops * Add insert_merge_block to block gen * Add block gen test for pos blocks * Start adding payloads to block gen * Fix clippy lints * Add execution payload to block gen * Add execute_payload to block_gen * Refactor block gen * Add all routes to mock server * Use Uint256 for base_fee_per_gas * Add working execution chain build * Remove unused var * Revert "Use Uint256 for base_fee_per_gas" This reverts commit 6c88f19ac45db834dd4dbf7a3c6e7242c1c0f735. * Fix base_fee_for_gas Uint256 * Update execute payload handle * Improve testing, fix bugs * Fix default fee-recipient * Fix fee-recipient address (again) * Add check for terminal block, add comments, tidy * Apply suggestions from code review Co-authored-by: realbigsean <seananderson33@GMAIL.com> * Fix is_none on handle Drop * Remove commented-out tests Co-authored-by: realbigsean <seananderson33@GMAIL.com>
2021-09-29 22:14:15 +00:00
}
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
Ok(())
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub async fn update_execution_engine_forkchoice(
self: &Arc<Self>,
current_slot: Slot,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
params: ForkchoiceUpdateParameters,
) -> Result<(), Error> {
let next_slot = current_slot + 1;
// There is no need to issue a `forkchoiceUpdated` (fcU) message unless the Bellatrix fork
// has:
//
// 1. Already happened.
// 2. Will happen in the next slot.
//
// The reason for a fcU message in the slot prior to the Bellatrix fork is in case the
// terminal difficulty has already been reached and a payload preparation message needs to
// be issued.
if self.slot_is_prior_to_bellatrix(next_slot) {
return Ok(());
}
let execution_layer = self
.execution_layer
.as_ref()
.ok_or(Error::ExecutionLayerMissing)?;
// Take the global lock for updating the execution engine fork choice.
//
// Whilst holding this lock we must:
//
// 1. Read the canonical head.
// 2. Issue a forkchoiceUpdated call to the execution engine.
//
// This will allow us to ensure that we provide the execution layer with an *ordered* view
// of the head. I.e., we will never communicate a past head after communicating a later
// one.
//
// There is a "deadlock warning" in this function. The downside of this nice ordering is the
// potential for deadlock. I would advise against any other use of
// `execution_engine_forkchoice_lock` apart from the one here.
let forkchoice_lock = execution_layer.execution_engine_forkchoice_lock().await;
let (head_block_root, head_hash, justified_hash, finalized_hash) = if let Some(head_hash) =
params.head_hash
{
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
(
params.head_root,
head_hash,
params
.justified_hash
.unwrap_or_else(ExecutionBlockHash::zero),
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
params
.finalized_hash
.unwrap_or_else(ExecutionBlockHash::zero),
)
} else {
// The head block does not have an execution block hash. We must check to see if we
// happen to be the proposer of the transition block, in which case we still need to
// send forkchoice_updated.
match self.spec.fork_name_at_slot::<T::EthSpec>(next_slot) {
// We are pre-bellatrix; no need to update the EL.
ForkName::Base | ForkName::Altair => return Ok(()),
_ => {
// We are post-bellatrix
if let Some(payload_attributes) = execution_layer
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
.payload_attributes(next_slot, params.head_root)
.await
{
// We are a proposer, check for terminal_pow_block_hash
if let Some(terminal_pow_block_hash) = execution_layer
.get_terminal_pow_block_hash(&self.spec, payload_attributes.timestamp)
.await
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
.map_err(Error::ForkchoiceUpdate)?
{
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
info!(
self.log,
"Prepared POS transition block proposer"; "slot" => next_slot
);
(
params.head_root,
terminal_pow_block_hash,
params
.justified_hash
.unwrap_or_else(ExecutionBlockHash::zero),
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
params
.finalized_hash
.unwrap_or_else(ExecutionBlockHash::zero),
)
} else {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// TTD hasn't been reached yet, no need to update the EL.
return Ok(());
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
} else {
// We are not a proposer, no need to update the EL.
return Ok(());
}
}
}
};
let forkchoice_updated_response = execution_layer
.notify_forkchoice_updated(
head_hash,
justified_hash,
finalized_hash,
current_slot,
head_block_root,
)
[Merge] Implement `execution_layer` (#2635) * Checkout serde_utils from rayonism * Make eth1::http functions pub * Add bones of execution_layer * Modify decoding * Expose Transaction, cargo fmt * Add executePayload * Add all minimal spec endpoints * Start adding json rpc wrapper * Finish custom JSON response handler * Switch to new rpc sending method * Add first test * Fix camelCase * Finish adding tests * Begin threading execution layer into BeaconChain * Fix clippy lints * Fix clippy lints * Thread execution layer into ClientBuilder * Add CLI flags * Add block processing methods to ExecutionLayer * Add block_on to execution_layer * Integrate execute_payload * Add extra_data field * Begin implementing payload handle * Send consensus valid/invalid messages * Fix minor type in task_executor * Call forkchoiceUpdated * Add search for TTD block * Thread TTD into execution layer * Allow producing block with execution payload * Add LRU cache for execution blocks * Remove duplicate 0x on ssz_types serialization * Add tests for block getter methods * Add basic block generator impl * Add is_valid_terminal_block to EL * Verify merge block in block_verification * Partially implement --terminal-block-hash-override * Add terminal_block_hash to ChainSpec * Remove Option from terminal_block_hash in EL * Revert merge changes to consensus/fork_choice * Remove commented-out code * Add bones for handling RPC methods on test server * Add first ExecutionLayer tests * Add testing for finding terminal block * Prevent infinite loops * Add insert_merge_block to block gen * Add block gen test for pos blocks * Start adding payloads to block gen * Fix clippy lints * Add execution payload to block gen * Add execute_payload to block_gen * Refactor block gen * Add all routes to mock server * Use Uint256 for base_fee_per_gas * Add working execution chain build * Remove unused var * Revert "Use Uint256 for base_fee_per_gas" This reverts commit 6c88f19ac45db834dd4dbf7a3c6e7242c1c0f735. * Fix base_fee_for_gas Uint256 * Update execute payload handle * Improve testing, fix bugs * Fix default fee-recipient * Fix fee-recipient address (again) * Add check for terminal block, add comments, tidy * Apply suggestions from code review Co-authored-by: realbigsean <seananderson33@GMAIL.com> * Fix is_none on handle Drop * Remove commented-out tests Co-authored-by: realbigsean <seananderson33@GMAIL.com>
2021-09-29 22:14:15 +00:00
.await
.map_err(Error::ExecutionForkChoiceUpdateFailed);
// The head has been read and the execution layer has been updated. It is now valid to send
// another fork choice update.
drop(forkchoice_lock);
match forkchoice_updated_response {
Ok(status) => match status {
PayloadStatus::Valid => {
// Ensure that fork choice knows that the block is no longer optimistic.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let chain = self.clone();
let fork_choice_update_result = self
.spawn_blocking_handle(
move || {
chain
.canonical_head
.fork_choice_write_lock()
.on_valid_execution_payload(head_block_root)
},
"update_execution_engine_valid_payload",
)
.await?;
if let Err(e) = fork_choice_update_result {
error!(
self.log,
"Failed to validate payload";
"error" => ?e
)
};
Ok(())
}
// There's nothing to be done for a syncing response. If the block is already
// `SYNCING` in fork choice, there's nothing to do. If already known to be `VALID`
// or `INVALID` then we don't want to change it to syncing.
PayloadStatus::Syncing => Ok(()),
// The specification doesn't list `ACCEPTED` as a valid response to a fork choice
// update. This response *seems* innocent enough, so we won't return early with an
// error. However, we create a log to bring attention to the issue.
PayloadStatus::Accepted => {
warn!(
self.log,
"Fork choice update received ACCEPTED";
"msg" => "execution engine provided an unexpected response to a fork \
choice update. although this is not a serious issue, please raise \
an issue."
);
Ok(())
}
PayloadStatus::Invalid {
latest_valid_hash,
ref validation_error,
} => {
debug!(
self.log,
"Invalid execution payload";
"validation_error" => ?validation_error,
"latest_valid_hash" => ?latest_valid_hash,
"head_hash" => ?head_hash,
"head_block_root" => ?head_block_root,
"method" => "fcU",
);
warn!(
self.log,
"Fork choice update invalidated payload";
"status" => ?status
);
// This implies that the terminal block was invalid. We are being explicit in
// invalidating only the head block in this case.
if latest_valid_hash == ExecutionBlockHash::zero() {
self.process_invalid_execution_payload(
&InvalidationOperation::InvalidateOne {
block_root: head_block_root,
},
)
.await?;
} else {
// The execution engine has stated that all blocks between the
// `head_execution_block_hash` and `latest_valid_hash` are invalid.
self.process_invalid_execution_payload(
&InvalidationOperation::InvalidateMany {
head_block_root,
always_invalidate_head: true,
latest_valid_ancestor: latest_valid_hash,
},
)
.await?;
}
Err(BeaconChainError::ExecutionForkChoiceUpdateInvalid { status })
}
PayloadStatus::InvalidBlockHash {
ref validation_error,
} => {
debug!(
self.log,
"Invalid execution payload block hash";
"validation_error" => ?validation_error,
"head_hash" => ?head_hash,
"head_block_root" => ?head_block_root,
"method" => "fcU",
);
warn!(
self.log,
"Fork choice update invalidated payload";
"status" => ?status
);
// The execution engine has stated that the head block is invalid, however it
// hasn't returned a latest valid ancestor.
//
// Using a `None` latest valid ancestor will result in only the head block
// being invalidated (no ancestors).
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
self.process_invalid_execution_payload(&InvalidationOperation::InvalidateOne {
block_root: head_block_root,
})
.await?;
Err(BeaconChainError::ExecutionForkChoiceUpdateInvalid { status })
}
},
Err(e) => Err(e),
}
[Merge] Implement `execution_layer` (#2635) * Checkout serde_utils from rayonism * Make eth1::http functions pub * Add bones of execution_layer * Modify decoding * Expose Transaction, cargo fmt * Add executePayload * Add all minimal spec endpoints * Start adding json rpc wrapper * Finish custom JSON response handler * Switch to new rpc sending method * Add first test * Fix camelCase * Finish adding tests * Begin threading execution layer into BeaconChain * Fix clippy lints * Fix clippy lints * Thread execution layer into ClientBuilder * Add CLI flags * Add block processing methods to ExecutionLayer * Add block_on to execution_layer * Integrate execute_payload * Add extra_data field * Begin implementing payload handle * Send consensus valid/invalid messages * Fix minor type in task_executor * Call forkchoiceUpdated * Add search for TTD block * Thread TTD into execution layer * Allow producing block with execution payload * Add LRU cache for execution blocks * Remove duplicate 0x on ssz_types serialization * Add tests for block getter methods * Add basic block generator impl * Add is_valid_terminal_block to EL * Verify merge block in block_verification * Partially implement --terminal-block-hash-override * Add terminal_block_hash to ChainSpec * Remove Option from terminal_block_hash in EL * Revert merge changes to consensus/fork_choice * Remove commented-out code * Add bones for handling RPC methods on test server * Add first ExecutionLayer tests * Add testing for finding terminal block * Prevent infinite loops * Add insert_merge_block to block gen * Add block gen test for pos blocks * Start adding payloads to block gen * Fix clippy lints * Add execution payload to block gen * Add execute_payload to block_gen * Refactor block gen * Add all routes to mock server * Use Uint256 for base_fee_per_gas * Add working execution chain build * Remove unused var * Revert "Use Uint256 for base_fee_per_gas" This reverts commit 6c88f19ac45db834dd4dbf7a3c6e7242c1c0f735. * Fix base_fee_for_gas Uint256 * Update execute payload handle * Improve testing, fix bugs * Fix default fee-recipient * Fix fee-recipient address (again) * Add check for terminal block, add comments, tidy * Apply suggestions from code review Co-authored-by: realbigsean <seananderson33@GMAIL.com> * Fix is_none on handle Drop * Remove commented-out tests Co-authored-by: realbigsean <seananderson33@GMAIL.com>
2021-09-29 22:14:15 +00:00
}
/// Returns `true` if the given slot is prior to the `bellatrix_fork_epoch`.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub fn slot_is_prior_to_bellatrix(&self, slot: Slot) -> bool {
self.spec.bellatrix_fork_epoch.map_or(true, |bellatrix| {
slot.epoch(T::EthSpec::slots_per_epoch()) < bellatrix
})
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
/// Returns the value of `execution_optimistic` for `block`.
///
/// Returns `Ok(false)` if the block is pre-Bellatrix, or has `ExecutionStatus::Valid`.
/// Returns `Ok(true)` if the block has `ExecutionStatus::Optimistic` or has
/// `ExecutionStatus::Invalid`.
pub fn is_optimistic_or_invalid_block<Payload: ExecPayload<T::EthSpec>>(
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
&self,
Add `execution_optimistic` flag to HTTP responses (#3070) ## Issue Addressed #3031 ## Proposed Changes Updates the following API endpoints to conform with https://github.com/ethereum/beacon-APIs/pull/190 and https://github.com/ethereum/beacon-APIs/pull/196 - [x] `beacon/states/{state_id}/root` - [x] `beacon/states/{state_id}/fork` - [x] `beacon/states/{state_id}/finality_checkpoints` - [x] `beacon/states/{state_id}/validators` - [x] `beacon/states/{state_id}/validators/{validator_id}` - [x] `beacon/states/{state_id}/validator_balances` - [x] `beacon/states/{state_id}/committees` - [x] `beacon/states/{state_id}/sync_committees` - [x] `beacon/headers` - [x] `beacon/headers/{block_id}` - [x] `beacon/blocks/{block_id}` - [x] `beacon/blocks/{block_id}/root` - [x] `beacon/blocks/{block_id}/attestations` - [x] `debug/beacon/states/{state_id}` - [x] `debug/beacon/heads` - [x] `validator/duties/attester/{epoch}` - [x] `validator/duties/proposer/{epoch}` - [x] `validator/duties/sync/{epoch}` Updates the following Server-Sent Events: - [x] `events?topics=head` - [x] `events?topics=block` - [x] `events?topics=finalized_checkpoint` - [x] `events?topics=chain_reorg` ## Backwards Incompatible There is a very minor breaking change with the way the API now handles requests to `beacon/blocks/{block_id}/root` and `beacon/states/{state_id}/root` when `block_id` or `state_id` is the `Root` variant of `BlockId` and `StateId` respectively. Previously a request to a non-existent root would simply echo the root back to the requester: ``` curl "http://localhost:5052/eth/v1/beacon/states/0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa/root" {"data":{"root":"0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"}} ``` Now it will return a `404`: ``` curl "http://localhost:5052/eth/v1/beacon/blocks/0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa/root" {"code":404,"message":"NOT_FOUND: beacon block with root 0xaaaa…aaaa","stacktraces":[]} ``` In addition to this is the block root `0x0000000000000000000000000000000000000000000000000000000000000000` previously would return the genesis block. It will now return a `404`: ``` curl "http://localhost:5052/eth/v1/beacon/blocks/0x0000000000000000000000000000000000000000000000000000000000000000" {"code":404,"message":"NOT_FOUND: beacon block with root 0x0000…0000","stacktraces":[]} ``` ## Additional Info - `execution_optimistic` is always set, and will return `false` pre-Bellatrix. I am also open to the idea of doing something like `#[serde(skip_serializing_if = "Option::is_none")]`. - The value of `execution_optimistic` is set to `false` where possible. Any computation that is reliant on the `head` will simply use the `ExecutionStatus` of the head (unless the head block is pre-Bellatrix). Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-07-25 08:23:00 +00:00
block: &SignedBeaconBlock<T::EthSpec, Payload>,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
) -> Result<bool, BeaconChainError> {
// Check if the block is pre-Bellatrix.
if self.slot_is_prior_to_bellatrix(block.slot()) {
Ok(false)
} else {
self.canonical_head
.fork_choice_read_lock()
.is_optimistic_or_invalid_block(&block.canonical_root())
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
.map_err(BeaconChainError::ForkChoiceError)
}
}
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
/// Returns the value of `execution_optimistic` for `head_block`.
///
/// Returns `Ok(false)` if the block is pre-Bellatrix, or has `ExecutionStatus::Valid`.
/// Returns `Ok(true)` if the block has `ExecutionStatus::Optimistic` or `ExecutionStatus::Invalid`.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
///
/// This function will return an error if `head_block` is not present in the fork choice store
/// and so should only be used on the head block or when the block *should* be present in the
/// fork choice store.
///
/// There is a potential race condition when syncing where the block_root of `head_block` could
/// be pruned from the fork choice store before being read.
pub fn is_optimistic_or_invalid_head_block<Payload: ExecPayload<T::EthSpec>>(
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
&self,
Add `execution_optimistic` flag to HTTP responses (#3070) ## Issue Addressed #3031 ## Proposed Changes Updates the following API endpoints to conform with https://github.com/ethereum/beacon-APIs/pull/190 and https://github.com/ethereum/beacon-APIs/pull/196 - [x] `beacon/states/{state_id}/root` - [x] `beacon/states/{state_id}/fork` - [x] `beacon/states/{state_id}/finality_checkpoints` - [x] `beacon/states/{state_id}/validators` - [x] `beacon/states/{state_id}/validators/{validator_id}` - [x] `beacon/states/{state_id}/validator_balances` - [x] `beacon/states/{state_id}/committees` - [x] `beacon/states/{state_id}/sync_committees` - [x] `beacon/headers` - [x] `beacon/headers/{block_id}` - [x] `beacon/blocks/{block_id}` - [x] `beacon/blocks/{block_id}/root` - [x] `beacon/blocks/{block_id}/attestations` - [x] `debug/beacon/states/{state_id}` - [x] `debug/beacon/heads` - [x] `validator/duties/attester/{epoch}` - [x] `validator/duties/proposer/{epoch}` - [x] `validator/duties/sync/{epoch}` Updates the following Server-Sent Events: - [x] `events?topics=head` - [x] `events?topics=block` - [x] `events?topics=finalized_checkpoint` - [x] `events?topics=chain_reorg` ## Backwards Incompatible There is a very minor breaking change with the way the API now handles requests to `beacon/blocks/{block_id}/root` and `beacon/states/{state_id}/root` when `block_id` or `state_id` is the `Root` variant of `BlockId` and `StateId` respectively. Previously a request to a non-existent root would simply echo the root back to the requester: ``` curl "http://localhost:5052/eth/v1/beacon/states/0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa/root" {"data":{"root":"0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"}} ``` Now it will return a `404`: ``` curl "http://localhost:5052/eth/v1/beacon/blocks/0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa/root" {"code":404,"message":"NOT_FOUND: beacon block with root 0xaaaa…aaaa","stacktraces":[]} ``` In addition to this is the block root `0x0000000000000000000000000000000000000000000000000000000000000000` previously would return the genesis block. It will now return a `404`: ``` curl "http://localhost:5052/eth/v1/beacon/blocks/0x0000000000000000000000000000000000000000000000000000000000000000" {"code":404,"message":"NOT_FOUND: beacon block with root 0x0000…0000","stacktraces":[]} ``` ## Additional Info - `execution_optimistic` is always set, and will return `false` pre-Bellatrix. I am also open to the idea of doing something like `#[serde(skip_serializing_if = "Option::is_none")]`. - The value of `execution_optimistic` is set to `false` where possible. Any computation that is reliant on the `head` will simply use the `ExecutionStatus` of the head (unless the head block is pre-Bellatrix). Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-07-25 08:23:00 +00:00
head_block: &SignedBeaconBlock<T::EthSpec, Payload>,
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
) -> Result<bool, BeaconChainError> {
// Check if the block is pre-Bellatrix.
if self.slot_is_prior_to_bellatrix(head_block.slot()) {
Ok(false)
} else {
self.canonical_head
.fork_choice_read_lock()
.is_optimistic_or_invalid_block_no_fallback(&head_block.canonical_root())
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
.map_err(BeaconChainError::ForkChoiceError)
}
}
/// Returns the value of `execution_optimistic` for the current head block.
/// You can optionally provide `head_info` if it was computed previously.
///
/// Returns `Ok(false)` if the head block is pre-Bellatrix, or has `ExecutionStatus::Valid`.
/// Returns `Ok(true)` if the head block has `ExecutionStatus::Optimistic` or `ExecutionStatus::Invalid`.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
///
/// There is a potential race condition when syncing where the block root of `head_info` could
/// be pruned from the fork choice store before being read.
pub fn is_optimistic_or_invalid_head(&self) -> Result<bool, BeaconChainError> {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
self.canonical_head
.head_execution_status()
.map(|status| status.is_optimistic_or_invalid())
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
}
pub fn is_optimistic_or_invalid_block_root(
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
&self,
block_slot: Slot,
block_root: &Hash256,
) -> Result<bool, BeaconChainError> {
// Check if the block is pre-Bellatrix.
if self.slot_is_prior_to_bellatrix(block_slot) {
Ok(false)
} else {
self.canonical_head
.fork_choice_read_lock()
.is_optimistic_or_invalid_block_no_fallback(block_root)
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
.map_err(BeaconChainError::ForkChoiceError)
}
}
/// This function takes a configured weak subjectivity `Checkpoint` and the latest finalized `Checkpoint`.
/// If the weak subjectivity checkpoint and finalized checkpoint share the same epoch, we compare
/// roots. If we the weak subjectivity checkpoint is from an older epoch, we iterate back through
/// roots in the canonical chain until we reach the finalized checkpoint from the correct epoch, and
/// compare roots. This must called on startup and during verification of any block which causes a finality
/// change affecting the weak subjectivity checkpoint.
pub fn verify_weak_subjectivity_checkpoint(
&self,
wss_checkpoint: Checkpoint,
beacon_block_root: Hash256,
state: &BeaconState<T::EthSpec>,
) -> Result<(), BeaconChainError> {
let finalized_checkpoint = state.finalized_checkpoint();
info!(self.log, "Verifying the configured weak subjectivity checkpoint"; "weak_subjectivity_epoch" => wss_checkpoint.epoch, "weak_subjectivity_root" => ?wss_checkpoint.root);
// If epochs match, simply compare roots.
if wss_checkpoint.epoch == finalized_checkpoint.epoch
&& wss_checkpoint.root != finalized_checkpoint.root
{
crit!(
self.log,
"Root found at the specified checkpoint differs";
"weak_subjectivity_root" => ?wss_checkpoint.root,
"finalized_checkpoint_root" => ?finalized_checkpoint.root
);
return Err(BeaconChainError::WeakSubjectivtyVerificationFailure);
} else if wss_checkpoint.epoch < finalized_checkpoint.epoch {
let slot = wss_checkpoint
.epoch
.start_slot(T::EthSpec::slots_per_epoch());
// Iterate backwards through block roots from the given state. If first slot of the epoch is a skip-slot,
// this will return the root of the closest prior non-skipped slot.
match self.root_at_slot_from_state(slot, beacon_block_root, state)? {
Some(root) => {
if root != wss_checkpoint.root {
crit!(
self.log,
"Root found at the specified checkpoint differs";
"weak_subjectivity_root" => ?wss_checkpoint.root,
"finalized_checkpoint_root" => ?finalized_checkpoint.root
);
return Err(BeaconChainError::WeakSubjectivtyVerificationFailure);
}
}
None => {
crit!(self.log, "The root at the start slot of the given epoch could not be found";
"wss_checkpoint_slot" => ?slot);
return Err(BeaconChainError::WeakSubjectivtyVerificationFailure);
}
}
}
Ok(())
}
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
/// Called by the timer on every slot.
///
/// Note: this function **MUST** be called from a non-async context since
/// it contains a call to `fork_choice` which may eventually call
/// `tokio::runtime::block_on` in certain cases.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
pub async fn per_slot_task(self: &Arc<Self>) {
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
if let Some(slot) = self.slot_clock.now() {
debug!(
self.log,
"Running beacon chain per slot tasks";
"slot" => ?slot
);
Avoid parallel fork choice runs during sync (#3217) ## Issue Addressed Fixes an issue that @paulhauner found with the v2.3.0 release candidate whereby the fork choice runs introduced by #3168 tripped over each other during sync: ``` May 24 23:06:40.542 WARN Error signalling fork choice waiter slot: 3884129, error: ForkChoiceSignalOutOfOrder { current: Slot(3884131), latest: Slot(3884129) }, service: beacon ``` This can occur because fork choice is called from the state advance _and_ the per-slot task. When one of these runs takes a long time it can end up finishing after a run from a later slot, tripping the error above. The problem is resolved by not running either of these fork choice calls during sync. Additionally, these parallel fork choice runs were causing issues in the database: ``` May 24 07:49:05.098 WARN Found a chain that should already have been pruned, head_slot: 92925, head_block_root: 0xa76c7bf1b98e54ed4b0d8686efcfdf853484e6c2a4c67e91cbf19e5ad1f96b17, service: beacon May 24 07:49:05.101 WARN Database migration failed error: HotColdDBError(FreezeSlotError { current_split_slot: Slot(92608), proposed_split_slot: Slot(92576) }), service: beacon ``` In this case, two fork choice calls triggering the finalization processing were being processed out of order due to differences in their processing time, causing the background migrator to try to advance finalization _backwards_ :flushed:. Removing the parallel fork choice runs from sync effectively addresses the issue, because these runs are most likely to have different finalized checkpoints (because of the speed at which fork choice advances during sync). In theory it's still possible to process updates out of order if any other fork choice runs end up completing out of order, but this should be much less common. Fixing out of order fork choice runs in general is difficult as it requires architectural changes like serialising fork choice updates through a single thread, or locking fork choice along with the head when it is mutated (https://github.com/sigp/lighthouse/pull/3175). ## Proposed Changes * Don't run per-slot fork choice during sync (if head is older than 4 slots) * Don't run state-advance fork choice during sync (if head is older than 4 slots) * Check for monotonic finalization updates in the background migrator. This is a good defensive check to have, and I'm not sure why we didn't have it before (we may have had it and wrongly removed it).
2022-05-25 03:27:30 +00:00
// Always run the light-weight pruning tasks (these structures should be empty during
// sync anyway).
self.naive_aggregation_pool.write().prune(slot);
self.block_times_cache.write().prune(slot);
// Don't run heavy-weight tasks during sync.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
if self.best_slot() + MAX_PER_SLOT_FORK_CHOICE_DISTANCE < slot {
Avoid parallel fork choice runs during sync (#3217) ## Issue Addressed Fixes an issue that @paulhauner found with the v2.3.0 release candidate whereby the fork choice runs introduced by #3168 tripped over each other during sync: ``` May 24 23:06:40.542 WARN Error signalling fork choice waiter slot: 3884129, error: ForkChoiceSignalOutOfOrder { current: Slot(3884131), latest: Slot(3884129) }, service: beacon ``` This can occur because fork choice is called from the state advance _and_ the per-slot task. When one of these runs takes a long time it can end up finishing after a run from a later slot, tripping the error above. The problem is resolved by not running either of these fork choice calls during sync. Additionally, these parallel fork choice runs were causing issues in the database: ``` May 24 07:49:05.098 WARN Found a chain that should already have been pruned, head_slot: 92925, head_block_root: 0xa76c7bf1b98e54ed4b0d8686efcfdf853484e6c2a4c67e91cbf19e5ad1f96b17, service: beacon May 24 07:49:05.101 WARN Database migration failed error: HotColdDBError(FreezeSlotError { current_split_slot: Slot(92608), proposed_split_slot: Slot(92576) }), service: beacon ``` In this case, two fork choice calls triggering the finalization processing were being processed out of order due to differences in their processing time, causing the background migrator to try to advance finalization _backwards_ :flushed:. Removing the parallel fork choice runs from sync effectively addresses the issue, because these runs are most likely to have different finalized checkpoints (because of the speed at which fork choice advances during sync). In theory it's still possible to process updates out of order if any other fork choice runs end up completing out of order, but this should be much less common. Fixing out of order fork choice runs in general is difficult as it requires architectural changes like serialising fork choice updates through a single thread, or locking fork choice along with the head when it is mutated (https://github.com/sigp/lighthouse/pull/3175). ## Proposed Changes * Don't run per-slot fork choice during sync (if head is older than 4 slots) * Don't run state-advance fork choice during sync (if head is older than 4 slots) * Check for monotonic finalization updates in the background migrator. This is a good defensive check to have, and I'm not sure why we didn't have it before (we may have had it and wrongly removed it).
2022-05-25 03:27:30 +00:00
return;
}
Run fork choice before block proposal (#3168) ## Issue Addressed Upcoming spec change https://github.com/ethereum/consensus-specs/pull/2878 ## Proposed Changes 1. Run fork choice at the start of every slot, and wait for this run to complete before proposing a block. 2. As an optimisation, also run fork choice 3/4 of the way through the slot (at 9s), _dequeueing attestations for the next slot_. 3. Remove the fork choice run from the state advance timer that occurred before advancing the state. ## Additional Info ### Block Proposal Accuracy This change makes us more likely to propose on top of the correct head in the presence of re-orgs with proposer boost in play. The main scenario that this change is designed to address is described in the linked spec issue. ### Attestation Accuracy This change _also_ makes us more likely to attest to the correct head. Currently in the case of a skipped slot at `slot` we only run fork choice 9s into `slot - 1`. This means the attestations from `slot - 1` aren't taken into consideration, and any boost applied to the block from `slot - 1` is not removed (it should be). In the language of the linked spec issue, this means we are liable to attest to C, even when the majority voting weight has already caused a re-org to B. ### Why remove the call before the state advance? If we've run fork choice at the start of the slot then it has already dequeued all the attestations from the previous slot, which are the only ones eligible to influence the head in the current slot. Running fork choice again is unnecessary (unless we run it for the next slot and try to pre-empt a re-org, but I don't currently think this is a great idea). ### Performance Based on Prater testing this adds about 5-25ms of runtime to block proposal times, which are 500-1000ms on average (and spike to 5s+ sometimes due to state handling issues :cry: ). I believe this is a small enough penalty to enable it by default, with the option to disable it via the new flag `--fork-choice-before-proposal-timeout 0`. Upcoming work on block packing and state representation will also reduce block production times in general, while removing the spikes. ### Implementation Fork choice gets invoked at the start of the slot via the `per_slot_task` function called from the slot timer. It then uses a condition variable to signal to block production that fork choice has been updated. This is a bit funky, but it seems to work. One downside of the timer-based approach is that it doesn't happen automatically in most of the tests. The test added by this PR has to trigger the run manually.
2022-05-20 05:02:11 +00:00
// Run fork choice and signal to any waiting task that it has completed.
self.recompute_head_at_current_slot().await;
Run fork choice before block proposal (#3168) ## Issue Addressed Upcoming spec change https://github.com/ethereum/consensus-specs/pull/2878 ## Proposed Changes 1. Run fork choice at the start of every slot, and wait for this run to complete before proposing a block. 2. As an optimisation, also run fork choice 3/4 of the way through the slot (at 9s), _dequeueing attestations for the next slot_. 3. Remove the fork choice run from the state advance timer that occurred before advancing the state. ## Additional Info ### Block Proposal Accuracy This change makes us more likely to propose on top of the correct head in the presence of re-orgs with proposer boost in play. The main scenario that this change is designed to address is described in the linked spec issue. ### Attestation Accuracy This change _also_ makes us more likely to attest to the correct head. Currently in the case of a skipped slot at `slot` we only run fork choice 9s into `slot - 1`. This means the attestations from `slot - 1` aren't taken into consideration, and any boost applied to the block from `slot - 1` is not removed (it should be). In the language of the linked spec issue, this means we are liable to attest to C, even when the majority voting weight has already caused a re-org to B. ### Why remove the call before the state advance? If we've run fork choice at the start of the slot then it has already dequeued all the attestations from the previous slot, which are the only ones eligible to influence the head in the current slot. Running fork choice again is unnecessary (unless we run it for the next slot and try to pre-empt a re-org, but I don't currently think this is a great idea). ### Performance Based on Prater testing this adds about 5-25ms of runtime to block proposal times, which are 500-1000ms on average (and spike to 5s+ sometimes due to state handling issues :cry: ). I believe this is a small enough penalty to enable it by default, with the option to disable it via the new flag `--fork-choice-before-proposal-timeout 0`. Upcoming work on block packing and state representation will also reduce block production times in general, while removing the spikes. ### Implementation Fork choice gets invoked at the start of the slot via the `per_slot_task` function called from the slot timer. It then uses a condition variable to signal to block production that fork choice has been updated. This is a bit funky, but it seems to work. One downside of the timer-based approach is that it doesn't happen automatically in most of the tests. The test added by this PR has to trigger the run manually.
2022-05-20 05:02:11 +00:00
// Send the notification regardless of fork choice success, this is a "best effort"
// notification and we don't want block production to hit the timeout in case of error.
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
// Use a blocking task to avoid blocking the core executor whilst waiting for locks
// in `ForkChoiceSignalTx`.
let chain = self.clone();
self.task_executor.clone().spawn_blocking(
move || {
// Signal block proposal for the next slot (if it happens to be waiting).
if let Some(tx) = &chain.fork_choice_signal_tx {
if let Err(e) = tx.notify_fork_choice_complete(slot) {
warn!(
chain.log,
"Error signalling fork choice waiter";
"error" => ?e,
"slot" => slot,
);
}
}
},
"per_slot_task_fc_signal_tx",
);
}
}
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
/// Runs the `map_fn` with the committee cache for `shuffling_epoch` from the chain with head
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
/// `head_block_root`. The `map_fn` will be supplied two values:
///
/// - `&CommitteeCache`: the committee cache that serves the given parameters.
/// - `Hash256`: the "shuffling decision root" which uniquely identifies the `CommitteeCache`.
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
///
/// It's not necessary that `head_block_root` matches our current view of the chain, it can be
/// any block that is:
///
/// - Known to us.
/// - The finalized block or a descendant of the finalized block.
///
/// It would be quite common for attestation verification operations to use a `head_block_root`
/// that differs from our view of the head.
///
/// ## Important
///
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
/// This function is **not** suitable for determining proposer duties (only attester duties).
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
///
/// ## Notes
///
/// This function exists in this odd "map" pattern because efficiently obtaining a committee
/// can be complex. It might involve reading straight from the `beacon_chain.shuffling_cache`
/// or it might involve reading it from a state from the DB. Due to the complexities of
/// `RwLock`s on the shuffling cache, a simple `Cow` isn't suitable here.
///
/// If the committee for `(head_block_root, shuffling_epoch)` isn't found in the
/// `shuffling_cache`, we will read a state from disk and then update the `shuffling_cache`.
Fix attestation shuffling filter (#3629) ## Issue Addressed Fix a bug in block production that results in blocks with 0 attestations during the first slot of an epoch. The bug is marked by debug logs of the form: > DEBG Discarding attestation because of missing ancestor, block_root: 0x3cc00d9c9e0883b2d0db8606278f2b8423d4902f9a1ee619258b5b60590e64f8, pivot_slot: 4042591 It occurs when trying to look up the shuffling decision root for an attestation from a slot which is prior to fork choice's finalized block. This happens frequently when proposing in the first slot of the epoch where we have: - `current_epoch == n` - `attestation.data.target.epoch == n - 1` - attestation shuffling epoch `== n - 3` (decision block being the last block of `n - 3`) - `state.finalized_checkpoint.epoch == n - 2` (first block of `n - 2` is finalized) Hence the shuffling decision slot is out of range of the fork choice backwards iterator _by a single slot_. Unfortunately this bug was hidden when we weren't pruning fork choice, and then reintroduced in v2.5.1 when we fixed the pruning (https://github.com/sigp/lighthouse/releases/tag/v2.5.1). There's no way to turn that off or disable the filtering in our current release, so we need a new release to fix this issue. Fortunately, it also does not occur on every epoch boundary because of the gradual pruning of fork choice every 256 blocks (~8 epochs): https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array_fork_choice.rs#L16 https://github.com/sigp/lighthouse/blob/01e84b71f524968f5b940fbd2fa31d29408b6581/consensus/proto_array/src/proto_array.rs#L713-L716 So the probability of proposing a 0-attestation block given a proposal assignment is approximately `1/32 * 1/8 = 0.39%`. ## Proposed Changes - Load the block's shuffling ID from fork choice and verify it against the expected shuffling ID of the head state. This code was initially written before we had settled on a representation of shuffling IDs, so I think it's a nice simplification to make use of them here rather than more ad-hoc logic that fundamentally does the same thing. ## Additional Info Thanks to @moshe-blox for noticing this issue and bringing it to our attention.
2022-10-18 04:02:06 +00:00
pub fn with_committee_cache<F, R>(
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
&self,
head_block_root: Hash256,
shuffling_epoch: Epoch,
map_fn: F,
) -> Result<R, Error>
where
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
F: Fn(&CommitteeCache, Hash256) -> Result<R, Error>,
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
{
let head_block = self
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
.canonical_head
.fork_choice_read_lock()
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
.get_block(&head_block_root)
.ok_or(Error::MissingBeaconBlock(head_block_root))?;
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
let shuffling_id = BlockShufflingIds {
current: head_block.current_epoch_shuffling_id.clone(),
next: head_block.next_epoch_shuffling_id.clone(),
block_root: head_block.root,
}
.id_for_epoch(shuffling_epoch)
.ok_or_else(|| Error::InvalidShufflingId {
shuffling_epoch,
head_block_epoch: head_block.slot.epoch(T::EthSpec::slots_per_epoch()),
})?;
// Obtain the shuffling cache, timing how long we wait.
let cache_wait_timer =
metrics::start_timer(&metrics::ATTESTATION_PROCESSING_SHUFFLING_CACHE_WAIT_TIMES);
let mut shuffling_cache = self
.shuffling_cache
.try_write_for(ATTESTATION_CACHE_LOCK_TIMEOUT)
.ok_or(Error::AttestationCacheLockTimeout)?;
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
metrics::stop_timer(cache_wait_timer);
Avoid duplicate committee cache loads (#3574) ## Issue Addressed NA ## Proposed Changes I have observed scenarios on Goerli where Lighthouse was receiving attestations which reference the same, un-cached shuffling on multiple threads at the same time. Lighthouse was then loading the same state from database and determining the shuffling on multiple threads at the same time. This is unnecessary load on the disk and RAM. This PR modifies the shuffling cache so that each entry can be either: - A committee - A promise for a committee (i.e., a `crossbeam_channel::Receiver`) Now, in the scenario where we have thread A and thread B simultaneously requesting the same un-cached shuffling, we will have the following: 1. Thread A will take the write-lock on the shuffling cache, find that there's no cached committee and then create a "promise" (a `crossbeam_channel::Sender`) for a committee before dropping the write-lock. 1. Thread B will then be allowed to take the write-lock for the shuffling cache and find the promise created by thread A. It will block the current thread waiting for thread A to fulfill that promise. 1. Thread A will load the state from disk, obtain the shuffling, send it down the channel, insert the entry into the cache and then continue to verify the attestation. 1. Thread B will then receive the shuffling from the receiver, be un-blocked and then continue to verify the attestation. In the case where thread A fails to generate the shuffling and drops the sender, the next time that specific shuffling is requested we will detect that the channel is disconnected and return a `None` entry for that shuffling. This will cause the shuffling to be re-calculated. ## Additional Info NA
2022-09-16 08:54:03 +00:00
if let Some(cache_item) = shuffling_cache.get(&shuffling_id) {
// The shuffling cache is no longer required, drop the write-lock to allow concurrent
// access.
drop(shuffling_cache);
Avoid duplicate committee cache loads (#3574) ## Issue Addressed NA ## Proposed Changes I have observed scenarios on Goerli where Lighthouse was receiving attestations which reference the same, un-cached shuffling on multiple threads at the same time. Lighthouse was then loading the same state from database and determining the shuffling on multiple threads at the same time. This is unnecessary load on the disk and RAM. This PR modifies the shuffling cache so that each entry can be either: - A committee - A promise for a committee (i.e., a `crossbeam_channel::Receiver`) Now, in the scenario where we have thread A and thread B simultaneously requesting the same un-cached shuffling, we will have the following: 1. Thread A will take the write-lock on the shuffling cache, find that there's no cached committee and then create a "promise" (a `crossbeam_channel::Sender`) for a committee before dropping the write-lock. 1. Thread B will then be allowed to take the write-lock for the shuffling cache and find the promise created by thread A. It will block the current thread waiting for thread A to fulfill that promise. 1. Thread A will load the state from disk, obtain the shuffling, send it down the channel, insert the entry into the cache and then continue to verify the attestation. 1. Thread B will then receive the shuffling from the receiver, be un-blocked and then continue to verify the attestation. In the case where thread A fails to generate the shuffling and drops the sender, the next time that specific shuffling is requested we will detect that the channel is disconnected and return a `None` entry for that shuffling. This will cause the shuffling to be re-calculated. ## Additional Info NA
2022-09-16 08:54:03 +00:00
let committee_cache = cache_item.wait()?;
map_fn(&committee_cache, shuffling_id.shuffling_decision_block)
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
} else {
Avoid duplicate committee cache loads (#3574) ## Issue Addressed NA ## Proposed Changes I have observed scenarios on Goerli where Lighthouse was receiving attestations which reference the same, un-cached shuffling on multiple threads at the same time. Lighthouse was then loading the same state from database and determining the shuffling on multiple threads at the same time. This is unnecessary load on the disk and RAM. This PR modifies the shuffling cache so that each entry can be either: - A committee - A promise for a committee (i.e., a `crossbeam_channel::Receiver`) Now, in the scenario where we have thread A and thread B simultaneously requesting the same un-cached shuffling, we will have the following: 1. Thread A will take the write-lock on the shuffling cache, find that there's no cached committee and then create a "promise" (a `crossbeam_channel::Sender`) for a committee before dropping the write-lock. 1. Thread B will then be allowed to take the write-lock for the shuffling cache and find the promise created by thread A. It will block the current thread waiting for thread A to fulfill that promise. 1. Thread A will load the state from disk, obtain the shuffling, send it down the channel, insert the entry into the cache and then continue to verify the attestation. 1. Thread B will then receive the shuffling from the receiver, be un-blocked and then continue to verify the attestation. In the case where thread A fails to generate the shuffling and drops the sender, the next time that specific shuffling is requested we will detect that the channel is disconnected and return a `None` entry for that shuffling. This will cause the shuffling to be re-calculated. ## Additional Info NA
2022-09-16 08:54:03 +00:00
// Create an entry in the cache that "promises" this value will eventually be computed.
// This avoids the case where multiple threads attempt to produce the same value at the
// same time.
//
// Creating the promise whilst we hold the `shuffling_cache` lock will prevent the same
// promise from being created twice.
let sender = shuffling_cache.create_promise(shuffling_id.clone())?;
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
// Drop the shuffling cache to avoid holding the lock for any longer than
// required.
drop(shuffling_cache);
debug!(
self.log,
"Committee cache miss";
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
"shuffling_id" => ?shuffling_epoch,
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
"head_block_root" => head_block_root.to_string(),
);
let state_read_timer =
metrics::start_timer(&metrics::ATTESTATION_PROCESSING_STATE_READ_TIMES);
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
// If the head of the chain can serve this request, use it.
//
// This code is a little awkward because we need to ensure that the head we read and
// the head we copy is identical. Taking one lock to read the head values and another
// to copy the head is liable to race-conditions.
let head_state_opt = self.with_head(|head| {
if head.beacon_block_root == head_block_root {
Ok(Some((
head.beacon_state
.clone_with(CloneConfig::committee_caches_only()),
head.beacon_state_root(),
)))
} else {
Ok::<_, Error>(None)
}
})?;
// If the head state is useful for this request, use it. Otherwise, read a state from
// disk.
let (mut state, state_root) = if let Some((state, state_root)) = head_state_opt {
(state, state_root)
} else {
let state_root = head_block.state_root;
let state = self
.store
.get_inconsistent_state_for_attestation_verification_only(
&state_root,
Some(head_block.slot),
)?
.ok_or(Error::MissingBeaconState(head_block.state_root))?;
(state, state_root)
};
/*
* IMPORTANT
*
* Since it's possible that
* `Store::get_inconsistent_state_for_attestation_verification_only` was used to obtain
* the state, we cannot rely upon the following fields:
*
* - `state.state_roots`
* - `state.block_roots`
*
* These fields should not be used for the rest of this function.
*/
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
metrics::stop_timer(state_read_timer);
let state_skip_timer =
metrics::start_timer(&metrics::ATTESTATION_PROCESSING_STATE_SKIP_TIMES);
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
// If the state is in an earlier epoch, advance it. If it's from a later epoch, reject
// it.
if state.current_epoch() + 1 < shuffling_epoch {
// Since there's a one-epoch look-ahead on the attester shuffling, it suffices to
// only advance into the slot prior to the `shuffling_epoch`.
let target_slot = shuffling_epoch
.saturating_sub(1_u64)
.start_slot(T::EthSpec::slots_per_epoch());
// Advance the state into the required slot, using the "partial" method since the state
// roots are not relevant for the shuffling.
partial_state_advance(&mut state, Some(state_root), target_slot, &self.spec)?;
} else if state.current_epoch() > shuffling_epoch {
return Err(Error::InvalidStateForShuffling {
state_epoch: state.current_epoch(),
shuffling_epoch,
});
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
}
metrics::stop_timer(state_skip_timer);
let committee_building_timer =
metrics::start_timer(&metrics::ATTESTATION_PROCESSING_COMMITTEE_BUILDING_TIMES);
let relative_epoch = RelativeEpoch::from_epoch(state.current_epoch(), shuffling_epoch)
.map_err(Error::IncorrectStateForAttestation)?;
state.build_committee_cache(relative_epoch, &self.spec)?;
Avoid duplicate committee cache loads (#3574) ## Issue Addressed NA ## Proposed Changes I have observed scenarios on Goerli where Lighthouse was receiving attestations which reference the same, un-cached shuffling on multiple threads at the same time. Lighthouse was then loading the same state from database and determining the shuffling on multiple threads at the same time. This is unnecessary load on the disk and RAM. This PR modifies the shuffling cache so that each entry can be either: - A committee - A promise for a committee (i.e., a `crossbeam_channel::Receiver`) Now, in the scenario where we have thread A and thread B simultaneously requesting the same un-cached shuffling, we will have the following: 1. Thread A will take the write-lock on the shuffling cache, find that there's no cached committee and then create a "promise" (a `crossbeam_channel::Sender`) for a committee before dropping the write-lock. 1. Thread B will then be allowed to take the write-lock for the shuffling cache and find the promise created by thread A. It will block the current thread waiting for thread A to fulfill that promise. 1. Thread A will load the state from disk, obtain the shuffling, send it down the channel, insert the entry into the cache and then continue to verify the attestation. 1. Thread B will then receive the shuffling from the receiver, be un-blocked and then continue to verify the attestation. In the case where thread A fails to generate the shuffling and drops the sender, the next time that specific shuffling is requested we will detect that the channel is disconnected and return a `None` entry for that shuffling. This will cause the shuffling to be re-calculated. ## Additional Info NA
2022-09-16 08:54:03 +00:00
let committee_cache = state.take_committee_cache(relative_epoch)?;
let committee_cache = Arc::new(committee_cache);
Optimize validator duties (#2243) ## Issue Addressed Closes #2052 ## Proposed Changes - Refactor the attester/proposer duties endpoints in the BN - Performance improvements - Fixes some potential inconsistencies with the dependent root fields. - Removes `http_api::beacon_proposer_cache` and just uses the one on the `BeaconChain` instead. - Move the code for the proposer/attester duties endpoints into separate files, for readability. - Refactor the `DutiesService` in the VC - Required to reduce the delay on broadcasting new blocks. - Gets rid of the `ValidatorDuty` shim struct that came about when we adopted the standard API. - Separate block/attestation duty tasks so that they don't block each other when one is slow. - In the VC, use `PublicKeyBytes` to represent validators instead of `PublicKey`. `PublicKey` is a legit crypto object whilst `PublicKeyBytes` is just a byte-array, it's much faster to clone/hash `PublicKeyBytes` and this change has had a significant impact on runtimes. - Unfortunately this has created lots of dust changes. - In the BN, store `PublicKeyBytes` in the `beacon_proposer_cache` and allow access to them. The HTTP API always sends `PublicKeyBytes` over the wire and the conversion from `PublicKey` -> `PublickeyBytes` is non-trivial, especially when queries have 100s/1000s of validators (like Pyrmont). - Add the `state_processing::state_advance` mod which dedups a lot of the "apply `n` skip slots to the state" code. - This also fixes a bug with some functions which were failing to include a state root as per [this comment](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/consensus/state_processing/src/state_advance.rs#L69-L74). I couldn't find any instance of this bug that resulted in anything more severe than keying a shuffling cache by the wrong block root. - Swap the VC block service to use `mpsc` from `tokio` instead of `futures`. This is consistent with the rest of the code base. ~~This PR *reduces* the size of the codebase :tada:~~ It *used* to reduce the size of the code base before I added more comments. ## Observations on Prymont - Proposer duties times down from peaks of 450ms to consistent <1ms. - Current epoch attester duties times down from >1s peaks to a consistent 20-30ms. - Block production down from +600ms to 100-200ms. ## Additional Info - ~~Blocked on #2241~~ - ~~Blocked on #2234~~ ## TODO - [x] ~~Refactor this into some smaller PRs?~~ Leaving this as-is for now. - [x] Address `per_slot_processing` roots. - [x] Investigate slow next epoch times. Not getting added to cache on block processing? - [x] Consider [this](https://github.com/sigp/lighthouse/blob/072695284f7eff82c51f79bc921ad942fea7483a/beacon_node/store/src/hot_cold_store.rs#L811-L812) in the scenario of replacing the state roots Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2021-03-17 05:09:57 +00:00
let shuffling_decision_block = shuffling_id.shuffling_decision_block;
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
self.shuffling_cache
.try_write_for(ATTESTATION_CACHE_LOCK_TIMEOUT)
.ok_or(Error::AttestationCacheLockTimeout)?
Avoid duplicate committee cache loads (#3574) ## Issue Addressed NA ## Proposed Changes I have observed scenarios on Goerli where Lighthouse was receiving attestations which reference the same, un-cached shuffling on multiple threads at the same time. Lighthouse was then loading the same state from database and determining the shuffling on multiple threads at the same time. This is unnecessary load on the disk and RAM. This PR modifies the shuffling cache so that each entry can be either: - A committee - A promise for a committee (i.e., a `crossbeam_channel::Receiver`) Now, in the scenario where we have thread A and thread B simultaneously requesting the same un-cached shuffling, we will have the following: 1. Thread A will take the write-lock on the shuffling cache, find that there's no cached committee and then create a "promise" (a `crossbeam_channel::Sender`) for a committee before dropping the write-lock. 1. Thread B will then be allowed to take the write-lock for the shuffling cache and find the promise created by thread A. It will block the current thread waiting for thread A to fulfill that promise. 1. Thread A will load the state from disk, obtain the shuffling, send it down the channel, insert the entry into the cache and then continue to verify the attestation. 1. Thread B will then receive the shuffling from the receiver, be un-blocked and then continue to verify the attestation. In the case where thread A fails to generate the shuffling and drops the sender, the next time that specific shuffling is requested we will detect that the channel is disconnected and return a `None` entry for that shuffling. This will cause the shuffling to be re-calculated. ## Additional Info NA
2022-09-16 08:54:03 +00:00
.insert_committee_cache(shuffling_id, &committee_cache);
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
metrics::stop_timer(committee_building_timer);
sender.send(committee_cache.clone());
Avoid duplicate committee cache loads (#3574) ## Issue Addressed NA ## Proposed Changes I have observed scenarios on Goerli where Lighthouse was receiving attestations which reference the same, un-cached shuffling on multiple threads at the same time. Lighthouse was then loading the same state from database and determining the shuffling on multiple threads at the same time. This is unnecessary load on the disk and RAM. This PR modifies the shuffling cache so that each entry can be either: - A committee - A promise for a committee (i.e., a `crossbeam_channel::Receiver`) Now, in the scenario where we have thread A and thread B simultaneously requesting the same un-cached shuffling, we will have the following: 1. Thread A will take the write-lock on the shuffling cache, find that there's no cached committee and then create a "promise" (a `crossbeam_channel::Sender`) for a committee before dropping the write-lock. 1. Thread B will then be allowed to take the write-lock for the shuffling cache and find the promise created by thread A. It will block the current thread waiting for thread A to fulfill that promise. 1. Thread A will load the state from disk, obtain the shuffling, send it down the channel, insert the entry into the cache and then continue to verify the attestation. 1. Thread B will then receive the shuffling from the receiver, be un-blocked and then continue to verify the attestation. In the case where thread A fails to generate the shuffling and drops the sender, the next time that specific shuffling is requested we will detect that the channel is disconnected and return a `None` entry for that shuffling. This will cause the shuffling to be re-calculated. ## Additional Info NA
2022-09-16 08:54:03 +00:00
map_fn(&committee_cache, shuffling_decision_block)
Implement standard eth2.0 API (#1569) - Resolves #1550 - Resolves #824 - Resolves #825 - Resolves #1131 - Resolves #1411 - Resolves #1256 - Resolve #1177 - Includes the `ShufflingId` struct initially defined in #1492. That PR is now closed and the changes are included here, with significant bug fixes. - Implement the https://github.com/ethereum/eth2.0-APIs in a new `http_api` crate using `warp`. This replaces the `rest_api` crate. - Add a new `common/eth2` crate which provides a wrapper around `reqwest`, providing the HTTP client that is used by the validator client and for testing. This replaces the `common/remote_beacon_node` crate. - Create a `http_metrics` crate which is a dedicated server for Prometheus metrics (they are no longer served on the same port as the REST API). We now have flags for `--metrics`, `--metrics-address`, etc. - Allow the `subnet_id` to be an optional parameter for `VerifiedUnaggregatedAttestation::verify`. This means it does not need to be provided unnecessarily by the validator client. - Move `fn map_attestation_committee` in `mod beacon_chain::attestation_verification` to a new `fn with_committee_cache` on the `BeaconChain` so the same cache can be used for obtaining validator duties. - Add some other helpers to `BeaconChain` to assist with common API duties (e.g., `block_root_at_slot`, `head_beacon_block_root`). - Change the `NaiveAggregationPool` so it can index attestations by `hash_tree_root(attestation.data)`. This is a requirement of the API. - Add functions to `BeaconChainHarness` to allow it to create slashings and exits. - Allow for `eth1::Eth1NetworkId` to go to/from a `String`. - Add functions to the `OperationPool` to allow getting all objects in the pool. - Add function to `BeaconState` to check if a committee cache is initialized. - Fix bug where `seconds_per_eth1_block` was not transferring over from `YamlConfig` to `ChainSpec`. - Add the `deposit_contract_address` to `YamlConfig` and `ChainSpec`. We needed to be able to return it in an API response. - Change some uses of serde `serialize_with` and `deserialize_with` to a single use of `with` (code quality). - Impl `Display` and `FromStr` for several BLS fields. - Check for clock discrepancy when VC polls BN for sync state (with +/- 1 slot tolerance). This is not intended to be comprehensive, it was just easy to do. - See #1434 for a per-endpoint overview. - Seeking clarity here: https://github.com/ethereum/eth2.0-APIs/issues/75 - [x] Add docs for prom port to close #1256 - [x] Follow up on this #1177 - [x] ~~Follow up with #1424~~ Will fix in future PR. - [x] Follow up with #1411 - [x] ~~Follow up with #1260~~ Will fix in future PR. - [x] Add quotes to all integers. - [x] Remove `rest_types` - [x] Address missing beacon block error. (#1629) - [x] ~~Add tests for lighthouse/peers endpoints~~ Wontfix - [x] ~~Follow up with validator status proposal~~ Tracked in #1434 - [x] Unify graffiti structs - [x] ~~Start server when waiting for genesis?~~ Will fix in future PR. - [x] TODO in http_api tests - [x] Move lighthouse endpoints off /eth/v1 - [x] Update docs to link to standard - ~~Blocked on #1586~~ Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-09-29 03:46:54 +00:00
}
}
/// Dumps the entire canonical chain, from the head to genesis to a vector for analysis.
///
/// This could be a very expensive operation and should only be done in testing/analysis
/// activities.
Separate execution payloads in the DB (#3157) ## Proposed Changes Reduce post-merge disk usage by not storing finalized execution payloads in Lighthouse's database. :warning: **This is achieved in a backwards-incompatible way for networks that have already merged** :warning:. Kiln users and shadow fork enjoyers will be unable to downgrade after running the code from this PR. The upgrade migration may take several minutes to run, and can't be aborted after it begins. The main changes are: - New column in the database called `ExecPayload`, keyed by beacon block root. - The `BeaconBlock` column now stores blinded blocks only. - Lots of places that previously used full blocks now use blinded blocks, e.g. analytics APIs, block replay in the DB, etc. - On finalization: - `prune_abanonded_forks` deletes non-canonical payloads whilst deleting non-canonical blocks. - `migrate_db` deletes finalized canonical payloads whilst deleting finalized states. - Conversions between blinded and full blocks are implemented in a compositional way, duplicating some work from Sean's PR #3134. - The execution layer has a new `get_payload_by_block_hash` method that reconstructs a payload using the EE's `eth_getBlockByHash` call. - I've tested manually that it works on Kiln, using Geth and Nethermind. - This isn't necessarily the most efficient method, and new engine APIs are being discussed to improve this: https://github.com/ethereum/execution-apis/pull/146. - We're depending on the `ethers` master branch, due to lots of recent changes. We're also using a workaround for https://github.com/gakonst/ethers-rs/issues/1134. - Payload reconstruction is used in the HTTP API via `BeaconChain::get_block`, which is now `async`. Due to the `async` fn, the `blocking_json` wrapper has been removed. - Payload reconstruction is used in network RPC to serve blocks-by-{root,range} responses. Here the `async` adjustment is messier, although I think I've managed to come up with a reasonable compromise: the handlers take the `SendOnDrop` by value so that they can drop it on _task completion_ (after the `fn` returns). Still, this is introducing disk reads onto core executor threads, which may have a negative performance impact (thoughts appreciated). ## Additional Info - [x] For performance it would be great to remove the cloning of full blocks when converting them to blinded blocks to write to disk. I'm going to experiment with a `put_block` API that takes the block by value, breaks it into a blinded block and a payload, stores the blinded block, and then re-assembles the full block for the caller. - [x] We should measure the latency of blocks-by-root and blocks-by-range responses. - [x] We should add integration tests that stress the payload reconstruction (basic tests done, issue for more extensive tests: https://github.com/sigp/lighthouse/issues/3159) - [x] We should (manually) test the schema v9 migration from several prior versions, particularly as blocks have changed on disk and some migrations rely on being able to load blocks. Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-05-12 00:42:17 +00:00
#[allow(clippy::type_complexity)]
pub fn chain_dump(
&self,
) -> Result<Vec<BeaconSnapshot<T::EthSpec, BlindedPayload<T::EthSpec>>>, Error> {
let mut dump = vec![];
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let mut last_slot = {
let head = self.canonical_head.cached_head();
BeaconSnapshot {
beacon_block: Arc::new(head.snapshot.beacon_block.clone_as_blinded()),
beacon_block_root: head.snapshot.beacon_block_root,
beacon_state: head.snapshot.beacon_state.clone(),
}
};
dump.push(last_slot.clone());
loop {
Update to Spec v0.10 (#817) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Disable/delete two outdated tests * Bump eth1 default vote warn to error * Delete outdated eth1 test Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-02-10 23:19:36 +00:00
let beacon_block_root = last_slot.beacon_block.parent_root();
Update to frozen spec ❄️ (v0.8.1) (#444) * types: first updates for v0.8 * state_processing: epoch processing v0.8.0 * state_processing: block processing v0.8.0 * tree_hash_derive: support generics in SignedRoot * types v0.8: update to use ssz_types * state_processing v0.8: use ssz_types * ssz_types: add bitwise methods and from_elem * types: fix v0.8 FIXMEs * ssz_types: add bitfield shift_up * ssz_types: iterators and DerefMut for VariableList * types,state_processing: use VariableList * ssz_types: fix BitVector Decode impl Fixed a typo in the implementation of ssz::Decode for BitVector, which caused it to be considered variable length! * types: fix test modules for v0.8 update * types: remove slow type-level arithmetic * state_processing: fix tests for v0.8 * op_pool: update for v0.8 * ssz_types: Bitfield difference length-independent Allow computing the difference of two bitfields of different lengths. * Implement compact committee support * epoch_processing: committee & active index roots * state_processing: genesis state builder v0.8 * state_processing: implement v0.8.1 * Further improve tree_hash * Strip examples, tests from cached_tree_hash * Update TreeHash, un-impl CachedTreeHash * Update bitfield TreeHash, un-impl CachedTreeHash * Update FixedLenVec TreeHash, unimpl CachedTreeHash * Update update tree_hash_derive for new TreeHash * Fix TreeHash, un-impl CachedTreeHash for ssz_types * Remove fixed_len_vec, ssz benches SSZ benches relied upon fixed_len_vec -- it is easier to just delete them and rebuild them later (when necessary) * Remove boolean_bitfield crate * Fix fake_crypto BLS compile errors * Update ef_tests for new v.8 type params * Update ef_tests submodule to v0.8.1 tag * Make fixes to support parsing ssz ef_tests * `compact_committee...` to `compact_committees...` * Derive more traits for `CompactCommittee` * Flip bitfield byte-endianness * Fix tree_hash for bitfields * Modify CLI output for ef_tests * Bump ssz crate version * Update ssz_types doc comment * Del cached tree hash tests from ssz_static tests * Tidy SSZ dependencies * Rename ssz_types crate to eth2_ssz_types * validator_client: update for v0.8 * ssz_types: update union/difference for bit order swap * beacon_node: update for v0.8, EthSpec * types: disable cached tree hash, update min spec * state_processing: fix slot bug in committee update * tests: temporarily disable fork choice harness test See #447 * committee cache: prevent out-of-bounds access In the case where we tried to access the committee of a shard that didn't have a committee in the current epoch, we were accessing elements beyond the end of the shuffling vector and panicking! This commit adds a check to make the failure safe and explicit. * fix bug in get_indexed_attestation and simplify There was a bug in our implementation of get_indexed_attestation whereby incorrect "committee indices" were used to index into the custody bitfield. The bug was only observable in the case where some bits of the custody bitfield were set to 1. The implementation has been simplified to remove the bug, and a test added. * state_proc: workaround for compact committees bug https://github.com/ethereum/eth2.0-specs/issues/1315 * v0.8: updates to make the EF tests pass * Remove redundant max operation checks. * Always supply both messages when checking attestation signatures -- allowing verification of an attestation with no signatures. * Swap the order of the fork and domain constant in `get_domain`, to match the spec. * rustfmt * ef_tests: add new epoch processing tests * Integrate v0.8 into master (compiles) * Remove unused crates, fix clippy lints * Replace v0.6.3 tags w/ v0.8.1 * Remove old comment * Ensure lmd ghost tests only run in release * Update readme
2019-07-30 02:44:51 +00:00
if beacon_block_root == Hash256::zero() {
break; // Genesis has been reached.
}
Separate execution payloads in the DB (#3157) ## Proposed Changes Reduce post-merge disk usage by not storing finalized execution payloads in Lighthouse's database. :warning: **This is achieved in a backwards-incompatible way for networks that have already merged** :warning:. Kiln users and shadow fork enjoyers will be unable to downgrade after running the code from this PR. The upgrade migration may take several minutes to run, and can't be aborted after it begins. The main changes are: - New column in the database called `ExecPayload`, keyed by beacon block root. - The `BeaconBlock` column now stores blinded blocks only. - Lots of places that previously used full blocks now use blinded blocks, e.g. analytics APIs, block replay in the DB, etc. - On finalization: - `prune_abanonded_forks` deletes non-canonical payloads whilst deleting non-canonical blocks. - `migrate_db` deletes finalized canonical payloads whilst deleting finalized states. - Conversions between blinded and full blocks are implemented in a compositional way, duplicating some work from Sean's PR #3134. - The execution layer has a new `get_payload_by_block_hash` method that reconstructs a payload using the EE's `eth_getBlockByHash` call. - I've tested manually that it works on Kiln, using Geth and Nethermind. - This isn't necessarily the most efficient method, and new engine APIs are being discussed to improve this: https://github.com/ethereum/execution-apis/pull/146. - We're depending on the `ethers` master branch, due to lots of recent changes. We're also using a workaround for https://github.com/gakonst/ethers-rs/issues/1134. - Payload reconstruction is used in the HTTP API via `BeaconChain::get_block`, which is now `async`. Due to the `async` fn, the `blocking_json` wrapper has been removed. - Payload reconstruction is used in network RPC to serve blocks-by-{root,range} responses. Here the `async` adjustment is messier, although I think I've managed to come up with a reasonable compromise: the handlers take the `SendOnDrop` by value so that they can drop it on _task completion_ (after the `fn` returns). Still, this is introducing disk reads onto core executor threads, which may have a negative performance impact (thoughts appreciated). ## Additional Info - [x] For performance it would be great to remove the cloning of full blocks when converting them to blinded blocks to write to disk. I'm going to experiment with a `put_block` API that takes the block by value, breaks it into a blinded block and a payload, stores the blinded block, and then re-assembles the full block for the caller. - [x] We should measure the latency of blocks-by-root and blocks-by-range responses. - [x] We should add integration tests that stress the payload reconstruction (basic tests done, issue for more extensive tests: https://github.com/sigp/lighthouse/issues/3159) - [x] We should (manually) test the schema v9 migration from several prior versions, particularly as blocks have changed on disk and some migrations rely on being able to load blocks. Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-05-12 00:42:17 +00:00
let beacon_block = self
.store
.get_blinded_block(&beacon_block_root)?
.ok_or_else(|| {
Error::DBInconsistent(format!("Missing block {}", beacon_block_root))
})?;
Update to Spec v0.10 (#817) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Disable/delete two outdated tests * Bump eth1 default vote warn to error * Delete outdated eth1 test Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-02-10 23:19:36 +00:00
let beacon_state_root = beacon_block.state_root();
let beacon_state = self
.store
Update to Spec v0.10 (#817) * Start updating types * WIP * Signature hacking * Existing EF tests passing with fake_crypto * Updates * Delete outdated API spec * The refactor continues * It compiles * WIP test fixes * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Cargo fmt * Fast aggregate verify test * Update REST API docs * Cargo fmt * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Address review comments; remove unused dependency * Disable/delete two outdated tests * Bump eth1 default vote warn to error * Delete outdated eth1 test Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
2020-02-10 23:19:36 +00:00
.get_state(&beacon_state_root, Some(beacon_block.slot()))?
.ok_or_else(|| {
Error::DBInconsistent(format!("Missing state {:?}", beacon_state_root))
})?;
Initial work towards v0.2.0 (#924) * Remove ping protocol * Initial renaming of network services * Correct rebasing relative to latest master * Start updating types * Adds HashMapDelay struct to utils * Initial network restructure * Network restructure. Adds new types for v0.2.0 * Removes build artefacts * Shift validation to beacon chain * Temporarily remove gossip validation This is to be updated to match current optimisation efforts. * Adds AggregateAndProof * Begin rebuilding pubsub encoding/decoding * Signature hacking * Shift gossipsup decoding into eth2_libp2p * Existing EF tests passing with fake_crypto * Shifts block encoding/decoding into RPC * Delete outdated API spec * All release tests passing bar genesis state parsing * Update and test YamlConfig * Update to spec v0.10 compatible BLS * Updates to BLS EF tests * Add EF test for AggregateVerify And delete unused hash2curve tests for uncompressed points * Update EF tests to v0.10.1 * Use optional block root correctly in block proc * Use genesis fork in deposit domain. All tests pass * Fast aggregate verify test * Update REST API docs * Fix unused import * Bump spec tags to v0.10.1 * Add `seconds_per_eth1_block` to chainspec * Update to timestamp based eth1 voting scheme * Return None from `get_votes_to_consider` if block cache is empty * Handle overflows in `is_candidate_block` * Revert to failing tests * Fix eth1 data sets test * Choose default vote according to spec * Fix collect_valid_votes tests * Fix `get_votes_to_consider` to choose all eligible blocks * Uncomment winning_vote tests * Add comments; remove unused code * Reduce seconds_per_eth1_block for simulation * Addressed review comments * Add test for default vote case * Fix logs * Remove unused functions * Meter default eth1 votes * Fix comments * Progress on attestation service * Address review comments; remove unused dependency * Initial work on removing libp2p lock * Add LRU caches to store (rollup) * Update attestation validation for DB changes (WIP) * Initial version of should_forward_block * Scaffold * Progress on attestation validation Also, consolidate prod+testing slot clocks so that they share much of the same implementation and can both handle sub-slot time changes. * Removes lock from libp2p service * Completed network lock removal * Finish(?) attestation processing * Correct network termination future * Add slot check to block check * Correct fmt issues * Remove Drop implementation for network service * Add first attempt at attestation proc. re-write * Add version 2 of attestation processing * Minor fixes * Add validator pubkey cache * Make get_indexed_attestation take a committee * Link signature processing into new attn verification * First working version * Ensure pubkey cache is updated * Add more metrics, slight optimizations * Clone committee cache during attestation processing * Update shuffling cache during block processing * Remove old commented-out code * Fix shuffling cache insert bug * Used indexed attestation in fork choice * Restructure attn processing, add metrics * Add more detailed metrics * Tidy, fix failing tests * Fix failing tests, tidy * Address reviewers suggestions * Disable/delete two outdated tests * Modification of validator for subscriptions * Add slot signing to validator client * Further progress on validation subscription * Adds necessary validator subscription functionality * Add new Pubkeys struct to signature_sets * Refactor with functional approach * Update beacon chain * Clean up validator <-> beacon node http types * Add aggregator status to ValidatorDuty * Impl Clone for manual slot clock * Fix minor errors * Further progress validator client subscription * Initial subscription and aggregation handling * Remove decompressed member from pubkey bytes * Progress to modifying val client for attestation aggregation * First draft of validator client upgrade for aggregate attestations * Add hashmap for indices lookup * Add state cache, remove store cache * Only build the head committee cache * Removes lock on a network channel * Partially implement beacon node subscription http api * Correct compilation issues * Change `get_attesting_indices` to use Vec * Fix failing test * Partial implementation of timer * Adds timer, removes exit_future, http api to op pool * Partial multiple aggregate attestation handling * Permits bulk messages accross gossipsub network channel * Correct compile issues * Improve gosispsub messaging and correct rest api helpers * Added global gossipsub subscriptions * Update validator subscriptions data structs * Tidy * Re-structure validator subscriptions * Initial handling of subscriptions * Re-structure network service * Add pubkey cache persistence file * Add more comments * Integrate persistence file into builder * Add pubkey cache tests * Add HashSetDelay and introduce into attestation service * Handles validator subscriptions * Add data_dir to beacon chain builder * Remove Option in pubkey cache persistence file * Ensure consistency between datadir/data_dir * Fix failing network test * Peer subnet discovery gets queued for future subscriptions * Reorganise attestation service functions * Initial wiring of attestation service * First draft of attestation service timing logic * Correct minor typos * Tidy * Fix todos * Improve tests * Add PeerInfo to connected peers mapping * Fix compile error * Fix compile error from merge * Split up block processing metrics * Tidy * Refactor get_pubkey_from_state * Remove commented-out code * Rename state_cache -> checkpoint_cache * Rename Checkpoint -> Snapshot * Tidy, add comments * Tidy up find_head function * Change some checkpoint -> snapshot * Add tests * Expose max_len * Remove dead code * Tidy * Fix bug * Add sync-speed metric * Add first attempt at VerifiableBlock * Start integrating into beacon chain * Integrate VerifiableBlock * Rename VerifableBlock -> PartialBlockVerification * Add start of typed methods * Add progress * Add further progress * Rename structs * Add full block verification to block_processing.rs * Further beacon chain integration * Update checks for gossip * Add todo * Start adding segement verification * Add passing chain segement test * Initial integration with batch sync * Minor changes * Tidy, add more error checking * Start adding chain_segment tests * Finish invalid signature tests * Include single and gossip verified blocks in tests * Add gossip verification tests * Start adding docs * Finish adding comments to block_processing.rs * Rename block_processing.rs -> block_verification * Start removing old block processing code * Fixes beacon_chain compilation * Fix project-wide compile errors * Remove old code * Correct code to pass all tests * Fix bug with beacon proposer index * Fix shim for BlockProcessingError * Only process one epoch at a time * Fix loop in chain segment processing * Correct tests from master merge * Add caching for state.eth1_data_votes * Add BeaconChain::validator_pubkey * Revert "Add caching for state.eth1_data_votes" This reverts commit cd73dcd6434fb8d8e6bf30c5356355598ea7b78e. Co-authored-by: Grant Wuerker <gwuerker@gmail.com> Co-authored-by: Michael Sproul <michael@sigmaprime.io> Co-authored-by: Michael Sproul <micsproul@gmail.com> Co-authored-by: pawan <pawandhananjay@gmail.com> Co-authored-by: Paul Hauner <paul@paulhauner.com>
2020-03-17 06:24:44 +00:00
let slot = BeaconSnapshot {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
beacon_block: Arc::new(beacon_block),
beacon_block_root,
beacon_state,
};
dump.push(slot.clone());
last_slot = slot;
}
dump.reverse();
Ok(dump)
}
/// Gets the current `EnrForkId`.
pub fn enr_fork_id(&self) -> EnrForkId {
// If we are unable to read the slot clock we assume that it is prior to genesis and
// therefore use the genesis slot.
let slot = self.slot().unwrap_or(self.spec.genesis_slot);
2020-04-08 06:46:37 +00:00
self.spec
.enr_fork_id::<T::EthSpec>(slot, self.genesis_validators_root)
}
/// Calculates the `Duration` to the next fork if it exists and returns it
/// with it's corresponding `ForkName`.
pub fn duration_to_next_fork(&self) -> Option<(ForkName, Duration)> {
// If we are unable to read the slot clock we assume that it is prior to genesis and
// therefore use the genesis slot.
let slot = self.slot().unwrap_or(self.spec.genesis_slot);
let (fork_name, epoch) = self.spec.next_fork_epoch::<T::EthSpec>(slot)?;
self.slot_clock
.duration_to_slot(epoch.start_slot(T::EthSpec::slots_per_epoch()))
.map(|duration| (fork_name, duration))
}
2020-04-21 15:05:46 +00:00
/// This method serves to get a sense of the current chain health. It is used in block proposal
/// to determine whether we should outsource payload production duties.
///
/// Since we are likely calling this during the slot we are going to propose in, don't take into
/// account the current slot when accounting for skips.
pub fn is_healthy(&self, parent_root: &Hash256) -> Result<ChainHealth, Error> {
// Check if the merge has been finalized.
if let Some(finalized_hash) = self
.canonical_head
.cached_head()
.forkchoice_update_parameters()
.finalized_hash
{
if ExecutionBlockHash::zero() == finalized_hash {
return Ok(ChainHealth::PreMerge);
}
} else {
return Ok(ChainHealth::PreMerge);
};
// Check that the parent is NOT optimistic.
if let Some(execution_status) = self
.canonical_head
.fork_choice_read_lock()
.get_block_execution_status(parent_root)
{
if execution_status.is_strictly_optimistic() {
return Ok(ChainHealth::Optimistic);
}
}
if self.config.builder_fallback_disable_checks {
return Ok(ChainHealth::Healthy);
}
let current_slot = self.slot()?;
// Check slots at the head of the chain.
let prev_slot = current_slot.saturating_sub(Slot::new(1));
let head_skips = prev_slot.saturating_sub(self.canonical_head.cached_head().head_slot());
let head_skips_check = head_skips.as_usize() <= self.config.builder_fallback_skips;
// Check if finalization is advancing.
let current_epoch = current_slot.epoch(T::EthSpec::slots_per_epoch());
let epochs_since_finalization = current_epoch.saturating_sub(
self.canonical_head
.cached_head()
.finalized_checkpoint()
.epoch,
);
let finalization_check = epochs_since_finalization.as_usize()
<= self.config.builder_fallback_epochs_since_finalization;
// Check skip slots in the last `SLOTS_PER_EPOCH`.
let start_slot = current_slot.saturating_sub(T::EthSpec::slots_per_epoch());
let mut epoch_skips = 0;
for slot in start_slot.as_u64()..current_slot.as_u64() {
if self
.block_root_at_slot_skips_none(Slot::new(slot))?
.is_none()
{
epoch_skips += 1;
}
}
let epoch_skips_check = epoch_skips <= self.config.builder_fallback_skips_per_epoch;
if !head_skips_check {
Ok(ChainHealth::Unhealthy(FailedCondition::Skips))
} else if !finalization_check {
Ok(ChainHealth::Unhealthy(
FailedCondition::EpochsSinceFinalization,
))
} else if !epoch_skips_check {
Ok(ChainHealth::Unhealthy(FailedCondition::SkipsPerEpoch))
} else {
Ok(ChainHealth::Healthy)
}
}
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
pub fn dump_as_dot<W: Write>(&self, output: &mut W) {
Use async code when interacting with EL (#3244) ## Overview This rather extensive PR achieves two primary goals: 1. Uses the finalized/justified checkpoints of fork choice (FC), rather than that of the head state. 2. Refactors fork choice, block production and block processing to `async` functions. Additionally, it achieves: - Concurrent forkchoice updates to the EL and cache pruning after a new head is selected. - Concurrent "block packing" (attestations, etc) and execution payload retrieval during block production. - Concurrent per-block-processing and execution payload verification during block processing. - The `Arc`-ification of `SignedBeaconBlock` during block processing (it's never mutated, so why not?): - I had to do this to deal with sending blocks into spawned tasks. - Previously we were cloning the beacon block at least 2 times during each block processing, these clones are either removed or turned into cheaper `Arc` clones. - We were also `Box`-ing and un-`Box`-ing beacon blocks as they moved throughout the networking crate. This is not a big deal, but it's nice to avoid shifting things between the stack and heap. - Avoids cloning *all the blocks* in *every chain segment* during sync. - It also has the potential to clean up our code where we need to pass an *owned* block around so we can send it back in the case of an error (I didn't do much of this, my PR is already big enough :sweat_smile:) - The `BeaconChain::HeadSafetyStatus` struct was removed. It was an old relic from prior merge specs. For motivation for this change, see https://github.com/sigp/lighthouse/pull/3244#issuecomment-1160963273 ## Changes to `canonical_head` and `fork_choice` Previously, the `BeaconChain` had two separate fields: ``` canonical_head: RwLock<Snapshot>, fork_choice: RwLock<BeaconForkChoice> ``` Now, we have grouped these values under a single struct: ``` canonical_head: CanonicalHead { cached_head: RwLock<Arc<Snapshot>>, fork_choice: RwLock<BeaconForkChoice> } ``` Apart from ergonomics, the only *actual* change here is wrapping the canonical head snapshot in an `Arc`. This means that we no longer need to hold the `cached_head` (`canonical_head`, in old terms) lock when we want to pull some values from it. This was done to avoid deadlock risks by preventing functions from acquiring (and holding) the `cached_head` and `fork_choice` locks simultaneously. ## Breaking Changes ### The `state` (root) field in the `finalized_checkpoint` SSE event Consider the scenario where epoch `n` is just finalized, but `start_slot(n)` is skipped. There are two state roots we might in the `finalized_checkpoint` SSE event: 1. The state root of the finalized block, which is `get_block(finalized_checkpoint.root).state_root`. 4. The state root at slot of `start_slot(n)`, which would be the state from (1), but "skipped forward" through any skip slots. Previously, Lighthouse would choose (2). However, we can see that when [Teku generates that event](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/beaconrestapi/src/main/java/tech/pegasys/teku/beaconrestapi/handlers/v1/events/EventSubscriptionManager.java#L171-L182) it uses [`getStateRootFromBlockRoot`](https://github.com/ConsenSys/teku/blob/de2b2801c89ef5abf983d6bf37867c37fc47121f/data/provider/src/main/java/tech/pegasys/teku/api/ChainDataProvider.java#L336-L341) which uses (1). I have switched Lighthouse from (2) to (1). I think it's a somewhat arbitrary choice between the two, where (1) is easier to compute and is consistent with Teku. ## Notes for Reviewers I've renamed `BeaconChain::fork_choice` to `BeaconChain::recompute_head`. Doing this helped ensure I broke all previous uses of fork choice and I also find it more descriptive. It describes an action and can't be confused with trying to get a reference to the `ForkChoice` struct. I've changed the ordering of SSE events when a block is received. It used to be `[block, finalized, head]` and now it's `[block, head, finalized]`. It was easier this way and I don't think we were making any promises about SSE event ordering so it's not "breaking". I've made it so fork choice will run when it's first constructed. I did this because I wanted to have a cached version of the last call to `get_head`. Ensuring `get_head` has been run *at least once* means that the cached values doesn't need to wrapped in an `Option`. This was fairly simple, it just involved passing a `slot` to the constructor so it knows *when* it's being run. When loading a fork choice from the store and a slot clock isn't handy I've just used the `slot` that was saved in the `fork_choice_store`. That seems like it would be a faithful representation of the slot when we saved it. I added the `genesis_time: u64` to the `BeaconChain`. It's small, constant and nice to have around. Since we're using FC for the fin/just checkpoints, we no longer get the `0x00..00` roots at genesis. You can see I had to remove a work-around in `ef-tests` here: b56be3bc2. I can't find any reason why this would be an issue, if anything I think it'll be better since the genesis-alias has caught us out a few times (0x00..00 isn't actually a real root). Edit: I did find a case where the `network` expected the 0x00..00 alias and patched it here: 3f26ac3e2. You'll notice a lot of changes in tests. Generally, tests should be functionally equivalent. Here are the things creating the most diff-noise in tests: - Changing tests to be `tokio::async` tests. - Adding `.await` to fork choice, block processing and block production functions. - Refactor of the `canonical_head` "API" provided by the `BeaconChain`. E.g., `chain.canonical_head.cached_head()` instead of `chain.canonical_head.read()`. - Wrapping `SignedBeaconBlock` in an `Arc`. - In the `beacon_chain/tests/block_verification`, we can't use the `lazy_static` `CHAIN_SEGMENT` variable anymore since it's generated with an async function. We just generate it in each test, not so efficient but hopefully insignificant. I had to disable `rayon` concurrent tests in the `fork_choice` tests. This is because the use of `rayon` and `block_on` was causing a panic. Co-authored-by: Mac L <mjladson@pm.me>
2022-07-03 05:36:50 +00:00
let canonical_head_hash = self.canonical_head.cached_head().head_block_root();
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
let mut visited: HashSet<Hash256> = HashSet::new();
let mut finalized_blocks: HashSet<Hash256> = HashSet::new();
let mut justified_blocks: HashSet<Hash256> = HashSet::new();
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
let genesis_block_hash = Hash256::zero();
writeln!(output, "digraph beacon {{").unwrap();
writeln!(output, "\t_{:?}[label=\"zero\"];", genesis_block_hash).unwrap();
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
// Canonical head needs to be processed first as otherwise finalized blocks aren't detected
// properly.
let heads = {
let mut heads = self.heads();
let canonical_head_index = heads
.iter()
.position(|(block_hash, _)| *block_hash == canonical_head_hash)
.unwrap();
let (canonical_head_hash, canonical_head_slot) =
heads.swap_remove(canonical_head_index);
heads.insert(0, (canonical_head_hash, canonical_head_slot));
heads
};
for (head_hash, _head_slot) in heads {
for maybe_pair in ParentRootBlockIterator::new(&*self.store, head_hash) {
let (block_hash, signed_beacon_block) = maybe_pair.unwrap();
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
if visited.contains(&block_hash) {
break;
}
visited.insert(block_hash);
if signed_beacon_block.slot() % T::EthSpec::slots_per_epoch() == 0 {
Separate execution payloads in the DB (#3157) ## Proposed Changes Reduce post-merge disk usage by not storing finalized execution payloads in Lighthouse's database. :warning: **This is achieved in a backwards-incompatible way for networks that have already merged** :warning:. Kiln users and shadow fork enjoyers will be unable to downgrade after running the code from this PR. The upgrade migration may take several minutes to run, and can't be aborted after it begins. The main changes are: - New column in the database called `ExecPayload`, keyed by beacon block root. - The `BeaconBlock` column now stores blinded blocks only. - Lots of places that previously used full blocks now use blinded blocks, e.g. analytics APIs, block replay in the DB, etc. - On finalization: - `prune_abanonded_forks` deletes non-canonical payloads whilst deleting non-canonical blocks. - `migrate_db` deletes finalized canonical payloads whilst deleting finalized states. - Conversions between blinded and full blocks are implemented in a compositional way, duplicating some work from Sean's PR #3134. - The execution layer has a new `get_payload_by_block_hash` method that reconstructs a payload using the EE's `eth_getBlockByHash` call. - I've tested manually that it works on Kiln, using Geth and Nethermind. - This isn't necessarily the most efficient method, and new engine APIs are being discussed to improve this: https://github.com/ethereum/execution-apis/pull/146. - We're depending on the `ethers` master branch, due to lots of recent changes. We're also using a workaround for https://github.com/gakonst/ethers-rs/issues/1134. - Payload reconstruction is used in the HTTP API via `BeaconChain::get_block`, which is now `async`. Due to the `async` fn, the `blocking_json` wrapper has been removed. - Payload reconstruction is used in network RPC to serve blocks-by-{root,range} responses. Here the `async` adjustment is messier, although I think I've managed to come up with a reasonable compromise: the handlers take the `SendOnDrop` by value so that they can drop it on _task completion_ (after the `fn` returns). Still, this is introducing disk reads onto core executor threads, which may have a negative performance impact (thoughts appreciated). ## Additional Info - [x] For performance it would be great to remove the cloning of full blocks when converting them to blinded blocks to write to disk. I'm going to experiment with a `put_block` API that takes the block by value, breaks it into a blinded block and a payload, stores the blinded block, and then re-assembles the full block for the caller. - [x] We should measure the latency of blocks-by-root and blocks-by-range responses. - [x] We should add integration tests that stress the payload reconstruction (basic tests done, issue for more extensive tests: https://github.com/sigp/lighthouse/issues/3159) - [x] We should (manually) test the schema v9 migration from several prior versions, particularly as blocks have changed on disk and some migrations rely on being able to load blocks. Co-authored-by: Paul Hauner <paul@paulhauner.com>
2022-05-12 00:42:17 +00:00
let block = self.get_blinded_block(&block_hash).unwrap().unwrap();
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
let state = self
.get_state(&block.state_root(), Some(block.slot()))
.unwrap()
.unwrap();
finalized_blocks.insert(state.finalized_checkpoint().root);
justified_blocks.insert(state.current_justified_checkpoint().root);
justified_blocks.insert(state.previous_justified_checkpoint().root);
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
}
if block_hash == canonical_head_hash {
writeln!(
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
output,
"\t_{:?}[label=\"{} ({})\" shape=box3d];",
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
block_hash,
block_hash,
signed_beacon_block.slot()
)
.unwrap();
} else if finalized_blocks.contains(&block_hash) {
writeln!(
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
output,
"\t_{:?}[label=\"{} ({})\" shape=Msquare];",
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
block_hash,
block_hash,
signed_beacon_block.slot()
)
.unwrap();
} else if justified_blocks.contains(&block_hash) {
writeln!(
output,
"\t_{:?}[label=\"{} ({})\" shape=cds];",
block_hash,
block_hash,
signed_beacon_block.slot()
)
.unwrap();
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
} else {
writeln!(
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
output,
"\t_{:?}[label=\"{} ({})\" shape=box];",
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
block_hash,
block_hash,
signed_beacon_block.slot()
)
.unwrap();
}
writeln!(
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
output,
"\t_{:?} -> _{:?};",
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
block_hash,
signed_beacon_block.parent_root()
)
.unwrap();
}
}
writeln!(output, "}}").unwrap();
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
}
/// Get a channel to request shutting down.
pub fn shutdown_sender(&self) -> Sender<ShutdownReason> {
self.shutdown_sender.clone()
}
Prune abandoned forks (#916) * Address compiler warning * Prune abandoned fork choice forks * New approach to pruning * Wrap some block hashes in a newtype pattern For increased type safety. * Add Graphviz chain dump emitter for debugging * Fix broken test case * Make prunes_abandoned_forks use real DiskStore * Mark finalized blocks in the GraphViz output * Refine debug stringification of Slot and Epoch Before this commit: print!("{:?}", Slot(123)) == "Slot(\n123\n)". After this commit: print!("{:?", Slot(123)) == "Slot(123)". * Simplify build_block() * Rewrite test case using more composable test primitives * Working rewritten test case * Tighten fork prunning test checks * Add another pruning test case * Bugfix: Finalized blocks weren't always properly detected * Pruning: Add pruning_does_not_touch_blocks_prior_to_finalization test case * Tighten pruning tests: check if heads are tracked properly * Add a failing test case for a buggy scenario * Change name of function to a more accurate one * Fix failing test case * Test case: Were skipped slots' states pruned? * Style fix: Simplify dereferencing * Tighten pruning tests: check if abandoned states are deleted * Towards atomicity of db ops * Correct typo * Prune also skipped slots' states * New logic for handling skipped states * Make skipped slots test pass * Post conflict resolution fixes * Formatting fixes * Tests passing * Block hashes in Graphviz node labels * Removed unused changes * Fix bug with states having < SlotsPerHistoricalRoot roots * Consolidate State/BlockRootsIterator for pruning * Address review feedback * Fix a bug in pruning tests * Detach prune_abandoned_forks() from its object * Move migrate.rs from store to beacon_chain * Move forks pruning onto a background thread * Bugfix: Heads weren't pruned when prune set contained only the head * Rename: freeze_to_state() -> process_finalization() * Eliminate redundant function parameter Co-authored-by: Michael Sproul <michael@sigmaprime.io>
2020-04-20 09:59:56 +00:00
// Used for debugging
#[allow(dead_code)]
pub fn dump_dot_file(&self, file_name: &str) {
let mut file = std::fs::File::create(file_name).unwrap();
self.dump_as_dot(&mut file);
}
/// Checks if attestations have been seen from the given `validator_index` at the
/// given `epoch`.
pub fn validator_seen_at_epoch(&self, validator_index: usize, epoch: Epoch) -> bool {
// It's necessary to assign these checks to intermediate variables to avoid a deadlock.
//
// See: https://github.com/sigp/lighthouse/pull/2230#discussion_r620013993
Ensure doppelganger detects attestations in blocks (#2495) ## Issue Addressed NA ## Proposed Changes When testing our (not-yet-released) Doppelganger implementation, I noticed that we aren't detecting attestations included in blocks (only those on the gossip network). This is because during [block processing](https://github.com/sigp/lighthouse/blob/e8c0d1f19b2736efb83c67a247e0022da5eaa7bb/beacon_node/beacon_chain/src/beacon_chain.rs#L2168) we only update the `observed_attestations` cache with each attestation, but not the `observed_attesters` cache. This is the correct behaviour when we consider the [p2p spec](https://github.com/ethereum/eth2.0-specs/blob/v1.0.1/specs/phase0/p2p-interface.md): > [IGNORE] There has been no other valid attestation seen on an attestation subnet that has an identical attestation.data.target.epoch and participating validator index. We're doing the right thing here and still allowing attestations on gossip that we've seen in a block. However, this doesn't work so nicely for Doppelganger. To resolve this, I've taken the following steps: - Add a `observed_block_attesters` cache. - Rename `observed_attesters` to `observed_gossip_attesters`. ## TODO - [x] Add a test to ensure a validator that's been seen in a block attestation (but not a gossip attestation) returns `true` for `BeaconChain::validator_seen_at_epoch`. - [x] Add a test to ensure `observed_block_attesters` isn't polluted via gossip attestations and vice versa. Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-08-09 02:43:03 +00:00
let gossip_attested = self
.observed_gossip_attesters
.read()
.index_seen_at_epoch(validator_index, epoch);
let block_attested = self
.observed_block_attesters
.read()
.index_seen_at_epoch(validator_index, epoch);
let aggregated = self
.observed_aggregators
.read()
.index_seen_at_epoch(validator_index, epoch);
let produced_block = self
.observed_block_producers
.read()
.index_seen_at_epoch(validator_index as u64, epoch);
Ensure doppelganger detects attestations in blocks (#2495) ## Issue Addressed NA ## Proposed Changes When testing our (not-yet-released) Doppelganger implementation, I noticed that we aren't detecting attestations included in blocks (only those on the gossip network). This is because during [block processing](https://github.com/sigp/lighthouse/blob/e8c0d1f19b2736efb83c67a247e0022da5eaa7bb/beacon_node/beacon_chain/src/beacon_chain.rs#L2168) we only update the `observed_attestations` cache with each attestation, but not the `observed_attesters` cache. This is the correct behaviour when we consider the [p2p spec](https://github.com/ethereum/eth2.0-specs/blob/v1.0.1/specs/phase0/p2p-interface.md): > [IGNORE] There has been no other valid attestation seen on an attestation subnet that has an identical attestation.data.target.epoch and participating validator index. We're doing the right thing here and still allowing attestations on gossip that we've seen in a block. However, this doesn't work so nicely for Doppelganger. To resolve this, I've taken the following steps: - Add a `observed_block_attesters` cache. - Rename `observed_attesters` to `observed_gossip_attesters`. ## TODO - [x] Add a test to ensure a validator that's been seen in a block attestation (but not a gossip attestation) returns `true` for `BeaconChain::validator_seen_at_epoch`. - [x] Add a test to ensure `observed_block_attesters` isn't polluted via gossip attestations and vice versa. Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-08-09 02:43:03 +00:00
gossip_attested || block_attested || aggregated || produced_block
}
}
Prepare for public testnet (#628) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Update to use v0.9.2 version of deposit contract * Add efforts to automate eth1 testnet deployment * Fix lcli testnet deployer * Modify bn CLI to parse eth2_testnet_dir * Progress with account_manager deposit tools * Make account manager submit deposits * Add password option for submitting deposits * Allow custom deposit amount * Add long names to lcli clap * Add password option to lcli deploy command * Fix minor bugs whilst testing * Address Michael's comments * Add refund-deposit-contract to lcli * Use time instead of skip count for denying long skips * Improve logging for eth1 * Fix bug with validator services exiting on error * Drop the block cache after genesis * Modify eth1 testnet config * Improve eth1 logging * Make validator wait until genesis time * Fix bug in eth1 voting * Add more logging to eth1 voting * Handle errors in eth1 http module * Set SECONDS_PER_DAY to sensible minimum * Shorten delay before testnet start * Ensure eth1 block is produced without any votes * Improve eth1 logging * Fix broken tests in eth1 * Tidy code in rest_api * Fix failing test in deposit_contract * Make CLI args more consistent * Change validator/duties endpoint * Add time-based skip slot limiting * Add new error type missed in previous commit * Add log when waiting for genesis * Refactor beacon node CLI * Remove unused dep * Add lcli eth1-genesis command * Fix bug in master merge * Apply clippy lints to beacon node * Add support for YamlConfig in Eth2TestnetDir * Upgrade tesnet deposit contract version * Remove unnecessary logging and correct formatting * Add a hardcoded eth2 testnet config * Ensure http server flag works. Overwrite configs with flags. * Ensure boot nodes are loaded from testnet dir * Fix account manager CLI bugs * Fix bugs with beacon node cli * Allow testnet dir without boot nodes * Write genesis state as SSZ * Remove ---/n from the start of testnet_dir files * Set default libp2p address * Tidy account manager CLI, add logging * Add check to see if testnet dir exists * Apply reviewers suggestions * Add HeadTracker struct * Add fork choice persistence * Shorten slot time for simulator * Add the /beacon/heads API endpoint * Update hardcoded testnet * Add tests for BeaconChain persistence + fix bugs * Extend BeaconChain persistence testing * Ensure chain is finalized b4 persistence tests * Ensure boot_enr.yaml is include in binary * Refactor beacon_chain_sim * Move files about in beacon sim * Update beacon_chain_sim * Fix bug with deposit inclusion * Increase log in genesis service, fix todo * Tidy sim, fix broken rest_api tests * Fix more broken tests * Update testnet * Fix broken rest api test * Tidy account manager CLI * Use tempdir for account manager * Stop hardcoded testnet dir from creating dir * Rename Eth2TestnetDir to Eth2TestnetConfig * Change hardcoded -> hard_coded * Tidy account manager * Add log to account manager * Tidy, ensure head tracker is loaded from disk * Tidy beacon chain builder * Tidy eth1_chain * Adds log support for simulator * Revert "Adds log support for simulator" This reverts commit ec77c66a052350f551db145cf20f213823428dd3. * Adds log support for simulator * Tidy after self-review * Change default log level * Address Michael's delicious PR comments * Fix off-by-one in tests
2019-12-03 04:28:57 +00:00
impl<T: BeaconChainTypes> Drop for BeaconChain<T> {
fn drop(&mut self) {
let drop = || -> Result<(), Error> {
self.persist_head_and_fork_choice()?;
self.persist_op_pool()?;
self.persist_eth1_cache()
};
if let Err(e) = drop() {
Prepare for public testnet (#628) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Update to use v0.9.2 version of deposit contract * Add efforts to automate eth1 testnet deployment * Fix lcli testnet deployer * Modify bn CLI to parse eth2_testnet_dir * Progress with account_manager deposit tools * Make account manager submit deposits * Add password option for submitting deposits * Allow custom deposit amount * Add long names to lcli clap * Add password option to lcli deploy command * Fix minor bugs whilst testing * Address Michael's comments * Add refund-deposit-contract to lcli * Use time instead of skip count for denying long skips * Improve logging for eth1 * Fix bug with validator services exiting on error * Drop the block cache after genesis * Modify eth1 testnet config * Improve eth1 logging * Make validator wait until genesis time * Fix bug in eth1 voting * Add more logging to eth1 voting * Handle errors in eth1 http module * Set SECONDS_PER_DAY to sensible minimum * Shorten delay before testnet start * Ensure eth1 block is produced without any votes * Improve eth1 logging * Fix broken tests in eth1 * Tidy code in rest_api * Fix failing test in deposit_contract * Make CLI args more consistent * Change validator/duties endpoint * Add time-based skip slot limiting * Add new error type missed in previous commit * Add log when waiting for genesis * Refactor beacon node CLI * Remove unused dep * Add lcli eth1-genesis command * Fix bug in master merge * Apply clippy lints to beacon node * Add support for YamlConfig in Eth2TestnetDir * Upgrade tesnet deposit contract version * Remove unnecessary logging and correct formatting * Add a hardcoded eth2 testnet config * Ensure http server flag works. Overwrite configs with flags. * Ensure boot nodes are loaded from testnet dir * Fix account manager CLI bugs * Fix bugs with beacon node cli * Allow testnet dir without boot nodes * Write genesis state as SSZ * Remove ---/n from the start of testnet_dir files * Set default libp2p address * Tidy account manager CLI, add logging * Add check to see if testnet dir exists * Apply reviewers suggestions * Add HeadTracker struct * Add fork choice persistence * Shorten slot time for simulator * Add the /beacon/heads API endpoint * Update hardcoded testnet * Add tests for BeaconChain persistence + fix bugs * Extend BeaconChain persistence testing * Ensure chain is finalized b4 persistence tests * Ensure boot_enr.yaml is include in binary * Refactor beacon_chain_sim * Move files about in beacon sim * Update beacon_chain_sim * Fix bug with deposit inclusion * Increase log in genesis service, fix todo * Tidy sim, fix broken rest_api tests * Fix more broken tests * Update testnet * Fix broken rest api test * Tidy account manager CLI * Use tempdir for account manager * Stop hardcoded testnet dir from creating dir * Rename Eth2TestnetDir to Eth2TestnetConfig * Change hardcoded -> hard_coded * Tidy account manager * Add log to account manager * Tidy, ensure head tracker is loaded from disk * Tidy beacon chain builder * Tidy eth1_chain * Adds log support for simulator * Revert "Adds log support for simulator" This reverts commit ec77c66a052350f551db145cf20f213823428dd3. * Adds log support for simulator * Tidy after self-review * Change default log level * Address Michael's delicious PR comments * Fix off-by-one in tests
2019-12-03 04:28:57 +00:00
error!(
self.log,
"Failed to persist on BeaconChain drop";
"error" => ?e
Prepare for public testnet (#628) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Update to use v0.9.2 version of deposit contract * Add efforts to automate eth1 testnet deployment * Fix lcli testnet deployer * Modify bn CLI to parse eth2_testnet_dir * Progress with account_manager deposit tools * Make account manager submit deposits * Add password option for submitting deposits * Allow custom deposit amount * Add long names to lcli clap * Add password option to lcli deploy command * Fix minor bugs whilst testing * Address Michael's comments * Add refund-deposit-contract to lcli * Use time instead of skip count for denying long skips * Improve logging for eth1 * Fix bug with validator services exiting on error * Drop the block cache after genesis * Modify eth1 testnet config * Improve eth1 logging * Make validator wait until genesis time * Fix bug in eth1 voting * Add more logging to eth1 voting * Handle errors in eth1 http module * Set SECONDS_PER_DAY to sensible minimum * Shorten delay before testnet start * Ensure eth1 block is produced without any votes * Improve eth1 logging * Fix broken tests in eth1 * Tidy code in rest_api * Fix failing test in deposit_contract * Make CLI args more consistent * Change validator/duties endpoint * Add time-based skip slot limiting * Add new error type missed in previous commit * Add log when waiting for genesis * Refactor beacon node CLI * Remove unused dep * Add lcli eth1-genesis command * Fix bug in master merge * Apply clippy lints to beacon node * Add support for YamlConfig in Eth2TestnetDir * Upgrade tesnet deposit contract version * Remove unnecessary logging and correct formatting * Add a hardcoded eth2 testnet config * Ensure http server flag works. Overwrite configs with flags. * Ensure boot nodes are loaded from testnet dir * Fix account manager CLI bugs * Fix bugs with beacon node cli * Allow testnet dir without boot nodes * Write genesis state as SSZ * Remove ---/n from the start of testnet_dir files * Set default libp2p address * Tidy account manager CLI, add logging * Add check to see if testnet dir exists * Apply reviewers suggestions * Add HeadTracker struct * Add fork choice persistence * Shorten slot time for simulator * Add the /beacon/heads API endpoint * Update hardcoded testnet * Add tests for BeaconChain persistence + fix bugs * Extend BeaconChain persistence testing * Ensure chain is finalized b4 persistence tests * Ensure boot_enr.yaml is include in binary * Refactor beacon_chain_sim * Move files about in beacon sim * Update beacon_chain_sim * Fix bug with deposit inclusion * Increase log in genesis service, fix todo * Tidy sim, fix broken rest_api tests * Fix more broken tests * Update testnet * Fix broken rest api test * Tidy account manager CLI * Use tempdir for account manager * Stop hardcoded testnet dir from creating dir * Rename Eth2TestnetDir to Eth2TestnetConfig * Change hardcoded -> hard_coded * Tidy account manager * Add log to account manager * Tidy, ensure head tracker is loaded from disk * Tidy beacon chain builder * Tidy eth1_chain * Adds log support for simulator * Revert "Adds log support for simulator" This reverts commit ec77c66a052350f551db145cf20f213823428dd3. * Adds log support for simulator * Tidy after self-review * Change default log level * Address Michael's delicious PR comments * Fix off-by-one in tests
2019-12-03 04:28:57 +00:00
)
} else {
info!(
self.log,
"Saved beacon chain to disk";
Prepare for public testnet (#628) * Update to spec v0.9.0 * Update to v0.9.1 * Bump spec tags for v0.9.1 * Formatting, fix CI failures * Resolve accidental KeyPair merge conflict * Document new BeaconState functions * Add `validator` changes from `validator-to-rest` * Add initial (failing) REST api tests * Fix signature parsing * Add more tests * Refactor http router * Add working tests for publish beacon block * Add validator duties tests * Move account_manager under `lighthouse` binary * Unify logfile handling in `environment` crate. * Fix incorrect cache drops in `advance_caches` * Update fork choice for v0.9.1 * Add `deposit_contract` crate * Add progress on validator onboarding * Add unfinished attesation code * Update account manager CLI * Write eth1 data file as hex string * Integrate ValidatorDirectory with validator_client * Move ValidatorDirectory into validator_client * Clean up some FIXMEs * Add beacon_chain_sim * Fix a few docs/logs * Expand `beacon_chain_sim` * Fix spec for `beacon_chain_sim * More testing for api * Start work on attestation endpoint * Reject empty attestations * Allow attestations to genesis block * Add working tests for `rest_api` validator endpoint * Remove grpc from beacon_node * Start heavy refactor of validator client - Block production is working * Prune old validator client files * Start works on attestation service * Add attestation service to validator client * Use full pubkey for validator directories * Add validator duties post endpoint * Use par_iter for keypair generation * Use bulk duties request in validator client * Add version http endpoint tests * Add interop keys and startup wait * Ensure a prompt exit * Add duties pruning * Fix compile error in beacon node tests * Add github workflow * Modify rust.yaml * Modify gitlab actions * Add to CI file * Add sudo to CI npm install * Move cargo fmt to own job in tests * Fix cargo fmt in CI * Add rustup update before cargo fmt * Change name of CI job * Make other CI jobs require cargo fmt * Add CI badge * Remove gitlab and travis files * Add different http timeout for debug * Update docker file, use makefile in CI * Use make in the dockerfile, skip the test * Use the makefile for debug GI test * Update book * Tidy grpc and misc things * Apply discv5 fixes * Address other minor issues * Fix warnings * Attempt fix for addr parsing * Tidy validator config, CLIs * Tidy comments * Tidy signing, reduce ForkService duplication * Fail if skipping too many slots * Set default recent genesis time to 0 * Add custom http timeout to validator * Fix compile bug in node_test_rig * Remove old bootstrap flag from val CLI * Update docs * Tidy val client * Change val client log levels * Add comments, more validity checks * Fix compile error, add comments * Undo changes to eth2-libp2p/src * Reduce duplication of keypair generation * Add more logging for validator duties * Fix beacon_chain_sim, nitpicks * Fix compile error, minor nits * Update to use v0.9.2 version of deposit contract * Add efforts to automate eth1 testnet deployment * Fix lcli testnet deployer * Modify bn CLI to parse eth2_testnet_dir * Progress with account_manager deposit tools * Make account manager submit deposits * Add password option for submitting deposits * Allow custom deposit amount * Add long names to lcli clap * Add password option to lcli deploy command * Fix minor bugs whilst testing * Address Michael's comments * Add refund-deposit-contract to lcli * Use time instead of skip count for denying long skips * Improve logging for eth1 * Fix bug with validator services exiting on error * Drop the block cache after genesis * Modify eth1 testnet config * Improve eth1 logging * Make validator wait until genesis time * Fix bug in eth1 voting * Add more logging to eth1 voting * Handle errors in eth1 http module * Set SECONDS_PER_DAY to sensible minimum * Shorten delay before testnet start * Ensure eth1 block is produced without any votes * Improve eth1 logging * Fix broken tests in eth1 * Tidy code in rest_api * Fix failing test in deposit_contract * Make CLI args more consistent * Change validator/duties endpoint * Add time-based skip slot limiting * Add new error type missed in previous commit * Add log when waiting for genesis * Refactor beacon node CLI * Remove unused dep * Add lcli eth1-genesis command * Fix bug in master merge * Apply clippy lints to beacon node * Add support for YamlConfig in Eth2TestnetDir * Upgrade tesnet deposit contract version * Remove unnecessary logging and correct formatting * Add a hardcoded eth2 testnet config * Ensure http server flag works. Overwrite configs with flags. * Ensure boot nodes are loaded from testnet dir * Fix account manager CLI bugs * Fix bugs with beacon node cli * Allow testnet dir without boot nodes * Write genesis state as SSZ * Remove ---/n from the start of testnet_dir files * Set default libp2p address * Tidy account manager CLI, add logging * Add check to see if testnet dir exists * Apply reviewers suggestions * Add HeadTracker struct * Add fork choice persistence * Shorten slot time for simulator * Add the /beacon/heads API endpoint * Update hardcoded testnet * Add tests for BeaconChain persistence + fix bugs * Extend BeaconChain persistence testing * Ensure chain is finalized b4 persistence tests * Ensure boot_enr.yaml is include in binary * Refactor beacon_chain_sim * Move files about in beacon sim * Update beacon_chain_sim * Fix bug with deposit inclusion * Increase log in genesis service, fix todo * Tidy sim, fix broken rest_api tests * Fix more broken tests * Update testnet * Fix broken rest api test * Tidy account manager CLI * Use tempdir for account manager * Stop hardcoded testnet dir from creating dir * Rename Eth2TestnetDir to Eth2TestnetConfig * Change hardcoded -> hard_coded * Tidy account manager * Add log to account manager * Tidy, ensure head tracker is loaded from disk * Tidy beacon chain builder * Tidy eth1_chain * Adds log support for simulator * Revert "Adds log support for simulator" This reverts commit ec77c66a052350f551db145cf20f213823428dd3. * Adds log support for simulator * Tidy after self-review * Change default log level * Address Michael's delicious PR comments * Fix off-by-one in tests
2019-12-03 04:28:57 +00:00
)
}
}
}
impl From<DBError> for Error {
fn from(e: DBError) -> Error {
2019-05-21 07:27:06 +00:00
Error::DBError(e)
}
}
impl From<ForkChoiceError> for Error {
fn from(e: ForkChoiceError) -> Error {
Error::ForkChoiceError(e)
}
}
impl From<BeaconStateError> for Error {
fn from(e: BeaconStateError) -> Error {
Error::BeaconStateError(e)
}
}
impl<T: EthSpec> ChainSegmentResult<T> {
pub fn into_block_error(self) -> Result<(), BlockError<T>> {
match self {
ChainSegmentResult::Failed { error, .. } => Err(error),
ChainSegmentResult::Successful { .. } => Ok(()),
}
}
}