Done in different PRs so that they can reviewed independently, as it's likely this won't be merged before I leave
Includes resolution for #4080
- [ ] #4299
- [ ] #4318
- [ ] #4320
Co-authored-by: Diva M <divma@protonmail.com>
Co-authored-by: Age Manning <Age@AgeManning.com>
## Issue Addressed
This PR addresses issue https://github.com/sigp/lighthouse/issues/4350
## Proposed Changes
This change will enable slasher broadcast in the following cases:
No flag is passed,
`--slasher-broadcast` is passed and,
`--slasher-broadcast=true` is passed.
Only when an explicit false value is passed the slasher does not broadcast.(`--slasher-broadcast=false`).
## Additional Info
TODO
- [x] Modify CLI parsing logic
- [x] Write test
Refer to #4353
Co-authored-by: Rahul Dogra <rahulcooldogra@gmail.com>
Co-authored-by: Gua00va <105484243+Gua00va@users.noreply.github.com>
## Issue Addressed
Closes#4354Closes#3987
Replaces #4305, #4283
## Proposed Changes
This switches the default slasher backend _back_ to LMDB.
If an MDBX database exists and the MDBX backend is enabled then MDBX will continue to be used. Our release binaries and Docker images will continue to include MDBX for as long as it is practical, so users of these should not notice any difference.
The main benefit is to users compiling from source and devs running tests. These users no longer have to struggle to compile MDBX and deal with the compatibility issues that arises. Similarly, devs don't need to worry about toggling feature flags in tests or risk forgetting to run the slasher tests due to backend issues.
## Issue Addressed
N/A
## Proposed Changes
This change will log the value of the relay block and the local block when the relay block is more profitable.
## Additional Info
This change will help validators understand the block selection (as it looks like the execution reward sometimes is higher that the MEV-reward).
The rationale for this change is to aid operators to better understand why a relay-block was chosen over a local block.
Looking at produced blocks (at beaconcha.in for example) it sometimes looks like the builder is making a profit just from the execution reward vs the MEV-reward, and creates the nagging question: "Could i have built this block and made that extra profit?"... The answer is probably "No, not without the extra transactions included by the relay", but by logging the value of the local block-candidate, this will no longer be an issue..
### Example (Mainnet)
https://beaconcha.in/block/17370329
MEV Block Reward: 0.17122 Ether to 0xE35bBaFa0266089f95d745d348b468622805D82B
Execution Reward: 0.17528 Ether to 0x1f9090aaE28b8a3dCeaDf281B0F12828e676c326
Difference: 0.00406 Ether
### Examples (Goerli)
https://goerli.beaconcha.in/block/9040065
MEV Block Reward: 0.56423 Ether to 0xF5794543CF6055Ae710E9c8E99E31343Cea004a8
Execution Reward: 0.56488 Ether to 0xfC0157aA4F5DB7177830ACddB3D5a9BB5BE9cc5e
Difference: 0.00065 Ether
https://goerli.beaconcha.in/block/9019921
MEV Block Reward: 1.39440 Ether to 0xF5794543CF6055Ae710E9c8E99E31343Cea004a8
Execution Reward: 1.39469 Ether to 0xfC0157aA4F5DB7177830ACddB3D5a9BB5BE9cc5e
Difference: 0.00029 Ether
https://goerli.beaconcha.in/block/9015583
MEV Block Reward: 1.04356 Ether to 0xF5794543CF6055Ae710E9c8E99E31343Cea004a8
Execution Reward: 1.04896 Ether to 0xfC0157aA4F5DB7177830ACddB3D5a9BB5BE9cc5e
Difference: 0.0054 Ether
## Issue Addressed
On deneb devnetv5, lighthouse keeps rate limiting peers which makes it harder to bootstrap new nodes as there are very few peers in the network. This PR adds an option to disable the inbound rate limiter for testnets.
Added an option to configure inbound rate limits as well.
Co-authored-by: Diva M <divma@protonmail.com>
## Proposed Changes
This is a light refactor of the execution layer's block hash calculation logic making it easier to use externally. e.g. in `eleel` (https://github.com/sigp/eleel/pull/18).
A static method is preferable to a method because the calculation doesn't actually need any data from `self`, and callers may want to compute block hashes without constructing an `ExecutionLayer` (`eleel` only constructs a simpler `Engine` struct).
This PR address the following spec change: https://github.com/ethereum/consensus-specs/pull/3312
Instead of subscribing to a long-lived subnet for every attached validator to a beacon node, all beacon nodes will subscribe to `SUBNETS_PER_NODE` long-lived subnets. This is currently set to 2 for mainnet.
This PR does not include any scoring or advanced discovery mechanisms. A future PR will improve discovery and we can implement scoring after the next hard fork when we expect all client teams and all implementations to respect this spec change.
This will be a significant change in the subnet network structure for consensus clients and we will likely have to monitor and tweak our peer management logic.
This PR adds the ability to read the Lighthouse logs from the HTTP API for both the BN and the VC.
This is done in such a way to as minimize any kind of performance hit by adding this feature.
The current design creates a tokio broadcast channel and mixes is into a form of slog drain that combines with our main global logger drain, only if the http api is enabled.
The drain gets the logs, checks the log level and drops them if they are below INFO. If they are INFO or higher, it sends them via a broadcast channel only if there are users subscribed to the HTTP API channel. If not, it drops the logs.
If there are more than one subscriber, the channel clones the log records and converts them to json in their independent HTTP API tasks.
Co-authored-by: Michael Sproul <micsproul@gmail.com>
## Issue Addressed
NA
## Proposed Changes
Adds metrics to track validators that are submitting equivocating (but not slashable) sync messages. This follows on from some research we've been doing in a separate fork of LH.
## Additional Info
@jimmygchen and @michaelsproul have already run their eyes over this so it should be easy to get into v4.2.0, IMO.
## Issue Addressed
#4281
## Proposed Changes
- Change `ShufflingCache` implementation from using `LruCache` to a custom cache that removes entry with lowest epoch instead of oldest insertion time.
- Protect the "enshrined" head shufflings when inserting new committee cache entries. The shuffling ids matching the head's previous, current, and future epochs will never be ejected from the cache during `Self::insert_cache_item`.
## Additional Info
There is a bonus point on shuffling preferences in the issue description that hasn't been implemented yet, as I haven't figured out a good way to do this:
> However I'm not convinced since there are some complexities around tie-breaking when two entries have the same epoch. Perhaps preferring entries in the canonical chain is best?
We should be able to check if a block is on the canonical chain by:
```rust
canonical_head
.fork_choice_read_lock()
.contains_block(root)
```
However we need to interleave the shuffling and fork choice locks, which may cause deadlocks if we're not careful (mentioned by @paulhauner). Alternatively, we could use the `state.block_roots` field of the `chain.canonical_head.snapshot.beacon_state`, which avoids deadlock but requires more work.
I'd like to get some feedback on review & testing before I dig deeper into the preferences stuff, as having the canonical head preference may already be quite useful in preventing the issue raised.
Co-authored-by: Jimmy Chen <jimmy@sigmaprime.io>
## Issue Addressed
Closes https://github.com/sigp/lighthouse/issues/4291, part of #3613.
## Proposed Changes
- Implement the `el_offline` field on `/eth/v1/node/syncing`. We set `el_offline=true` if:
- The EL's internal status is `Offline` or `AuthFailed`, _or_
- The most recent call to `newPayload` resulted in an error (more on this in a moment).
- Use the `el_offline` field in the VC to mark nodes with offline ELs as _unsynced_. These nodes will still be used, but only after synced nodes.
- Overhaul the usage of `RequireSynced` so that `::No` is used almost everywhere. The `--allow-unsynced` flag was broken and had the opposite effect to intended, so it has been deprecated.
- Add tests for the EL being offline on the upcheck call, and being offline due to the newPayload check.
## Why track `newPayload` errors?
Tracking the EL's online/offline status is too coarse-grained to be useful in practice, because:
- If the EL is timing out to some calls, it's unlikely to timeout on the `upcheck` call, which is _just_ `eth_syncing`. Every failed call is followed by an upcheck [here](693886b941/beacon_node/execution_layer/src/engines.rs (L372-L380)), which would have the effect of masking the failure and keeping the status _online_.
- The `newPayload` call is the most likely to time out. It's the call in which ELs tend to do most of their work (often 1-2 seconds), with `forkchoiceUpdated` usually returning much faster (<50ms).
- If `newPayload` is failing consistently (e.g. timing out) then this is a good indication that either the node's EL is in trouble, or the network as a whole is. In the first case validator clients _should_ prefer other BNs if they have one available. In the second case, all of their BNs will likely report `el_offline` and they'll just have to proceed with trying to use them.
## Additional Changes
- Add utility method `ForkName::latest` which is quite convenient for test writing, but probably other things too.
- Delete some stale comments from when we used to support multiple execution nodes.
## Issue Addressed
#2335
## Proposed Changes
- Remove the `lighthouse-network::tests::gossipsub_tests` module
- Remove dead code from the `lighthouse-network::tests::common` helper module (`build_full_mesh`)
## Additional Info
After discussion with both @divagant-martian and @AgeManning, these tests seem to have two main issues in that they are:
- Redundant, in that they don't test anything meaningful (due to our handling of duplicate messages)
- Out-of-place, in that it doesn't really test Lighthouse-specific functionality (rather libp2p functionality)
As such, this PR supersedes #4286.
## Issue Addressed
NA
## Proposed Changes
Adds an additional check to a feature introduced in #4179 to prevent us from re-queuing already-known blocks that could be rejected immediately.
## Additional Info
Ideally this would have been included in v4.1.0, however we came across it too late to release it safely. We decided that the safest path forward is to release *without* this check and then patch it in the next version. The lack of this check should only result in a very minor performance impact (the impact is totally negligible in my assessment).
## Issue Addressed
NA
## Proposed Changes
Adds a flag to store invalid blocks on disk for teh debugz. Only *some* invalid blocks are stored, those which:
- Were received via gossip (rather than RPC, for instance)
- This keeps things simple to start with and should capture most blocks.
- Passed gossip verification
- This reduces the ability for random people to fill up our disk. A proposer signature is required to write something to disk.
## Additional Info
It's possible that we'll store blocks that aren't necessarily invalid, but we had an internal error during verification. Those blocks seem like they might be useful sometimes.
## Issue Addressed
N/A
## Proposed Changes
Replace ganache-cli with anvil https://github.com/foundry-rs/foundry/blob/master/anvil/README.md
We can lose all js dependencies in CI as a consequence.
## Additional info
Also changes the ethers-rs version used in the execution layer (for the transaction reconstruction) to a newer one. This was necessary to get use the ethers utils for anvil. The fixed execution engine integration tests should catch any potential issues with the payload reconstruction after #3592
Co-authored-by: Michael Sproul <michael@sigmaprime.io>
## Issue Addressed
#4233
## Proposed Changes
Remove the `best_justified_checkpoint` from the `PersistedForkChoiceStore` type as it is now unused.
Additionally, remove the `Option`'s wrapping the `justified_checkpoint` and `finalized_checkpoint` fields on `ProtoNode` which were only present to facilitate a previous migration.
Include the necessary code to facilitate the migration to a new DB schema.
## Issue Addressed
Addresses #4238
## Proposed Changes
- [x] Add tests for the scenarios
- [x] Use the fork of the attestation slot for signature verification.
## Issue Addressed
Addresses #4234
## Proposed Changes
- Skip withdrawals processing in an inconsistent state replay.
- Repurpose `StateRootStrategy`: rename to `StateProcessingStrategy` and always skip withdrawals if using `StateProcessingStrategy::Inconsistent`
- Add a test to reproduce the scenario
Co-authored-by: Jimmy Chen <jimmy@sigmaprime.io>
## Issue Addressed
#4266
## Proposed Changes
- Log `Using external block builder` instead of `Connected to external block builder` on its initialization to resolve the confusion (there's no actual connection there)
## Additional Info
The log is mentioned in builders docs, so it's changed there too.
This commit adds a check to the networking service when handling core gossipsub topic subscription requests. If the BN is already subscribed to the core topics, we won't attempt to resubscribe.
## Issue Addressed
#4258
## Proposed Changes
- In the networking service, check if we're already subscribed to all of the core gossipsub topics and, if so, do nothing
## Additional Info
N/A
## Limit Backfill Sync
This PR transitions Lighthouse from syncing all the way back to genesis to only syncing back to the weak subjectivity point (~ 5 months) when syncing via a checkpoint sync.
There are a number of important points to note with this PR:
- Firstly and most importantly, this PR fundamentally shifts the default security guarantees of checkpoint syncing in Lighthouse. Prior to this PR, Lighthouse could verify the checkpoint of any given chain by ensuring the chain eventually terminates at the corresponding genesis. This guarantee can still be employed via the new CLI flag --genesis-backfill which will prompt lighthouse to the old behaviour of downloading all blocks back to genesis. The new behaviour only checks the proposer signatures for the last 5 months of blocks but cannot guarantee the chain matches the genesis chain.
- I have not modified any of the peer scoring or RPC responses. Clients syncing from gensis, will downscore new Lighthouse peers that do not possess blocks prior to the WSP. This is by design, as Lighthouse nodes of this form, need a mechanism to sort through peers in order to find useful peers in order to complete their genesis sync. We therefore do not discriminate between empty/error responses for blocks prior or post the local WSP. If we request a block that a peer does not posses, then fundamentally that peer is less useful to us than other peers.
- This will make a radical shift in that the majority of nodes will no longer store the full history of the chain. In the future we could add a pruning mechanism to remove old blocks from the db also.
Co-authored-by: Paul Hauner <paul@paulhauner.com>
## Issue Addressed
#3873
## Proposed Changes
add a cache to optimise historical state lookup.
## Additional Info
N/A
Co-authored-by: Michael Sproul <micsproul@gmail.com>
## Issue Addressed
This PR un-deprecates some commonly used test util functions, e.g. `extend_chain`. Most of these were deprecated in 2020 but some of us still found them quite convenient and they're still being used a lot. If there's no issue with using them, I think we should remove the "Deprecated" comment to avoid confusion.
## Issue Addressed
#4150
## Proposed Changes
Maintain trusted peers in the pruning logic. ~~In principle the changes here are not necessary as a trusted peer has a max score (100) and all other peers can have at most 0 (because we don't implement positive scores). This means that we should never prune trusted peers unless we have more trusted peers than the target peer count.~~
This change shifts this logic to explicitly never prune trusted peers which I expect is the intuitive behaviour.
~~I suspect the issue in #4150 arises when a trusted peer disconnects from us for one reason or another and then we remove that peer from our peerdb as it becomes stale. When it re-connects at some large time later, it is no longer a trusted peer.~~
Currently we do disconnect trusted peers, and this PR corrects this to maintain trusted peers in the pruning logic.
As suggested in #4150 we maintain trusted peers in the db and thus we remember them even if they disconnect from us.
It is a well-known fact that IP addresses for beacon nodes used by specific validators can be de-anonymized. There is an assumed risk that a malicious user may attempt to DOS validators when producing blocks to prevent chain growth/liveness.
Although there are a number of ideas put forward to address this, there a few simple approaches we can take to mitigate this risk.
Currently, a Lighthouse user is able to set a number of beacon-nodes that their validator client can connect to. If one beacon node is taken offline, it can fallback to another. Different beacon nodes can use VPNs or rotate IPs in order to mask their IPs.
This PR provides an additional setup option which further mitigates attacks of this kind.
This PR introduces a CLI flag --proposer-only to the beacon node. Setting this flag will configure the beacon node to run with minimal peers and crucially will not subscribe to subnets or sync committees. Therefore nodes of this kind should not be identified as nodes connected to validators of any kind.
It also introduces a CLI flag --proposer-nodes to the validator client. Users can then provide a number of beacon nodes (which may or may not run the --proposer-only flag) that the Validator client will use for block production and propagation only. If these nodes fail, the validator client will fallback to the default list of beacon nodes.
Users are then able to set up a number of beacon nodes dedicated to block proposals (which are unlikely to be identified as validator nodes) and point their validator clients to produce blocks on these nodes and attest on other beacon nodes. An attack attempting to prevent liveness on the eth2 network would then need to preemptively find and attack the proposer nodes which is significantly more difficult than the default setup.
This is a follow on from: #3328
Co-authored-by: Michael Sproul <michael@sigmaprime.io>
Co-authored-by: Paul Hauner <paul@paulhauner.com>
## Issue Addressed
NA
## Proposed Changes
Avoids reprocessing loops introduced in #4179. (Also somewhat related to #4192).
Breaks the re-queue loop by only re-queuing when an RPC block is received before the attestation creation deadline.
I've put `proposal_is_known` behind a closure to avoid interacting with the `observed_proposers` lock unnecessarily.
## Additional Info
NA
## Issue Addressed
Closes#4185
## Proposed Changes
- Set user agent to `Lighthouse/vX.Y.Z-<commit hash>` by default
- Allow tweaking user agent via `--builder-user-agent "agent"`
## Proposed Changes
Builds on #4028 to use the new payload bodies methods in the HTTP API as well.
## Caveats
The payloads by range method only works for the finalized chain, so it can't be used in the execution engine integration tests because we try to reconstruct unfinalized payloads there.
## Issue Addressed
NA
## Proposed Changes
Apply two changes to code introduced in #4179:
1. Remove the `ERRO` log for when we error on `proposer_has_been_observed()`. We were seeing a lot of this in our logs for finalized blocks and it's a bit noisy.
1. Use `false` rather than `true` for `proposal_already_known` when there is an error. If a block raises an error in `proposer_has_been_observed()` then the block must be invalid, so we should process (and reject) it now rather than queuing it.
For reference, here is one of the offending `ERRO` logs:
```
ERRO Failed to check observed proposers block_root: 0x5845…878e, source: rpc, error: FinalizedBlock { slot: Slot(5410983), finalized_slot: Slot(5411232) }
```
## Additional Info
NA
## Issue Addressed
NA
## Proposed Changes
Similar to #4181 but without the version bump and a more nuanced fix.
Patches the high CPU usage seen after the Capella fork which was caused by processing exits when there are skip slots.
## Additional Info
~~This is an imperfect solution that will cause us to drop some exits at the fork boundary. This is tracked at #4184.~~
## Proposed Changes
We already make some attempts to avoid processing RPC blocks when a block from the same proposer is already being processed through gossip. This PR strengthens that guarantee by using the existing cache for `observed_block_producers` to inform whether an RPC block's processing should be delayed.
## Proposed Changes
This change attempts to prevent failed re-orgs by:
1. Lowering the re-org cutoff from 2s to 1s. This is informed by a failed re-org attempted by @yorickdowne's node. The failed block was requested in the 1.5-2s window due to a Vouch failure, and failed to propagate to the majority of the network before the attestation deadline at 4s.
2. Allow users to adjust their re-org cutoff depending on observed network conditions and their risk profile. The static 2 second cutoff was too rigid.
3. Add a `--proposer-reorg-disallowed-offsets` flag which can be used to prohibit reorgs at certain slots. This is intended to help workaround an issue whereby reorging blocks at slot 1 are currently taking ~1.6s to propagate on gossip rather than ~500ms. This is suspected to be due to a cache miss in current versions of Prysm, which should be fixed in their next release.
## Additional Info
I'm of two minds about removing the `shuffling_stable` check which checks for blocks at slot 0 in the epoch. If we removed it users would be able to configure Lighthouse to try reorging at slot 0, which likely wouldn't work very well due to interactions with the proposer index cache. I think we could leave it for now and revisit it later.
## Issue Addressed
#4146
## Proposed Changes
Removes the `ExecutionOptimisticForkVersionedResponse` type and the associated Beacon API endpoint which is now deprecated. Also removes the test associated with the endpoint.
## Issue Addressed
N/A
## Proposed Changes
Adds a flag for disabling peer scoring. This is useful for local testing and testing small networks for new features.
> This is currently a WIP and all features are subject to alteration or removal at any time.
## Overview
The successor to #2873.
Contains the backbone of `beacon.watch` including syncing code, the initial API, and several core database tables.
See `watch/README.md` for more information, requirements and usage.
It is possible that when we go to ban a peer, there is already an unbanned message in the queue. It could lead to the case that we ban and immediately unban a peer leaving us in a state where a should-be banned peer is unbanned.
If this banned peer connects to us in this faulty state, we currently do not attempt to re-ban it. This PR does correct this also, so if we do see this error, it will now self-correct (although we shouldn't see the error in the first place).
I have also incremented the severity of not supporting protocols as I see peers ultimately get banned in a few steps and it seems to make sense to just ban them outright, rather than have them linger.
## Issue Addressed
#3212
## Proposed Changes
- Introduce a new `rate_limiting_backfill_queue` - any new inbound backfill work events gets immediately sent to this FIFO queue **without any processing**
- Spawn a `backfill_scheduler` routine that pops a backfill event from the FIFO queue at specified intervals (currently halfway through a slot, or at 6s after slot start for 12s slots) and sends the event to `BeaconProcessor` via a `scheduled_backfill_work_tx` channel
- This channel gets polled last in the `InboundEvents`, and work event received is wrapped in a `InboundEvent::ScheduledBackfillWork` enum variant, which gets processed immediately or queued by the `BeaconProcessor` (existing logic applies from here)
Diagram comparing backfill processing with / without rate-limiting:
https://github.com/sigp/lighthouse/issues/3212#issuecomment-1386249922
See this comment for @paulhauner's explanation and solution: https://github.com/sigp/lighthouse/issues/3212#issuecomment-1384674956
## Additional Info
I've compared this branch (with backfill processing rate limited to to 1 and 3 batches per slot) against the latest stable version. The CPU usage during backfill sync is reduced by ~5% - 20%, more details on this page:
https://hackmd.io/@jimmygchen/SJuVpJL3j
The above testing is done on Goerli (as I don't currently have hardware for Mainnet), I'm guessing the differences are likely to be bigger on mainnet due to block size.
### TODOs
- [x] Experiment with processing multiple batches per slot. (need to think about how to do this for different slot durations)
- [x] Add option to disable rate-limiting, enabed by default.
- [x] (No longer required now we're reusing the reprocessing queue) Complete the `backfill_scheduler` task when backfill sync is completed or not required
