Commit Graph

4376 Commits

Author SHA1 Message Date
ethDreamer
52e5083502
Fixed bugs for m3 readiness (#2669)
* Fixed bugs for m3 readiness

* woops

* cargo fmt..
2021-12-02 14:26:53 +11:00
Paul Hauner
b162b067de
Misc changes for merge testnets (#2667)
* Thread eth1_block_hash into interop genesis state

* Add merge-fork-epoch flag

* Build LH with minimal spec by default

* Add verbose logs to execution_layer

* Add --http-allow-sync-stalled flag

* Update lcli new-testnet to create genesis state

* Fix http test

* Fix compile errors in tests
2021-12-02 14:26:52 +11:00
Paul Hauner
a1033a9247
Add BeaconChainHarness tests for The Merge (#2661)
* Start adding merge tests

* Expose MockExecutionLayer

* Add mock_execution_layer to BeaconChainHarness

* Progress with merge test

* Return more detailed errors with gas limit issues

* Use a better gas limit in block gen

* Ensure TTD is met in block gen

* Fix basic_merge tests

* Start geth testing

* Fix conflicts after rebase

* Remove geth tests

* Improve merge test

* Address clippy lints

* Make pow block gen a pure function

* Add working new test, breaking existing test

* Fix test names

* Add should_panic

* Don't run merge tests in debug

* Detect a tokio runtime when starting MockServer

* Fix clippy lint, include merge tests
2021-12-02 14:26:52 +11:00
Paul Hauner
801f6f7425
Disable autotests for beacon_chain (#2658) 2021-12-02 14:26:52 +11:00
Pawan Dhananjay
471cf10392
Add merge fork_epoch and fork_version to Config (#2663) 2021-12-02 14:26:52 +11:00
realbigsean
5efb7aeab4
add automated docker build for merge-f2f branch (#2654) 2021-12-02 14:26:52 +11:00
Paul Hauner
01031931d9
[Merge] Add execution API test vectors from Geth (#2651)
* Add geth request vectors

* Add geth response vectors

* Fix clippy lints
2021-12-02 14:26:52 +11:00
Paul Hauner
20ca7a56ed
[Merge] Add serde impls for Transactions type (#2649)
* Start implemented serde for transactions

* Revise serde impl

* Add tests for transaction decoding
2021-12-02 14:26:51 +11:00
Paul Hauner
d8623cfc4f
[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>
2021-12-02 14:26:51 +11:00
ethDreamer
1563bce905
Finished Gossip Block Validation Conditions (#2640)
* Gossip Block Validation is Much More Efficient

Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-12-02 14:26:51 +11:00
ethDreamer
fe945bc84a
Fork boundary fix (#2646)
* Fixed Gossip Topics on Fork Boundary
2021-12-02 14:26:51 +11:00
realbigsean
aa534f8989
Store execution block hash in fork choice (#2643)
* - Update the fork choice `ProtoNode` to include `is_merge_complete`
- Add database migration for the persisted fork choice

* update tests

* Small cleanup

* lints

* store execution block hash in fork choice rather than bool
2021-12-02 14:26:51 +11:00
Michael Sproul
fd828199f5
Update test vectors to v1.1.0 (#2642) 2021-12-02 14:26:51 +11:00
Paul Hauner
18eee2dc82
Handle merge fork in web3signer (#2631) 2021-12-02 14:26:51 +11:00
Michael Sproul
10b263fed4
Update merge consensus to v1.1.0-beta.5 (#2630) 2021-12-02 14:26:50 +11:00
Paul Hauner
c10e8ce955
Fix clippy lints on merge-f2f (#2626)
* Remove unchecked arith from ssz_derive

* Address clippy lints in block_verfication

* Use safe math for is_valid_gas_limit
2021-12-02 14:26:50 +11:00
Michael Sproul
cce855f9ea
Fix consensus, SSZ, tree hash & run merge EF tests (#2622)
* Update to v1.1.0-beta.4 (squash of #2548)

* SSZ, cached tree hash, EF tests
2021-12-02 14:26:50 +11:00
Mark Mackey
5687c56d51
Initial merge changes
Added Execution Payload from Rayonism Fork

Updated new Containers to match Merge Spec

Updated BeaconBlockBody for Merge Spec

Completed updating BeaconState and BeaconBlockBody

Modified ExecutionPayload<T> to use Transaction<T>

Mostly Finished Changes for beacon-chain.md

Added some things for fork-choice.md

Update to match new fork-choice.md/fork.md changes

ran cargo fmt

Added Missing Pieces in eth2_libp2p for Merge

fix ef test

Various Changes to Conform Closer to Merge Spec
2021-12-02 14:26:50 +11:00
Mac L
fe75a0a9a1 Add background file logging (#2762)
## Issue Addressed

Closes #1996 

## Proposed Changes

Run a second `Logger` via `sloggers` which logs to a file in the background with:
- separate `debug-level` for background and terminal logging
- the ability to limit log size
- rotation through a customizable number of log files
- an option to compress old log files (`.gz` format)

Add the following new CLI flags:
- `--logfile-debug-level`: The debug level of the log files
- `--logfile-max-size`: The maximum size of each log file
- `--logfile-max-number`: The number of old log files to store
- `--logfile-compress`: Whether to compress old log files

By default background logging uses the `debug` log level and saves logfiles to:
- Beacon Node:  `$HOME/.lighthouse/$network/beacon/logs/beacon.log`
- Validator Client:  `$HOME/.lighthouse/$network/validators/logs/validator.log`

Or, when using the `--datadir` flag:
`$datadir/beacon/logs/beacon.log` and `$datadir/validators/logs/validator.log`

Once rotated, old logs are stored like so: `beacon.log.1`, `beacon.log.2` etc. 
> Note: `beacon.log.1` is always newer than `beacon.log.2`.

## Additional Info

Currently the default value of `--logfile-max-size` is 200 (MB) and `--logfile-max-number` is 5.
This means that the maximum storage space that the logs will take up by default is 1.2GB. 
(200MB x 5 from old log files + <200MB the current logfile being written to)
Happy to adjust these default values to whatever people think is appropriate. 

It's also worth noting that when logging to a file, we lose our custom `slog` formatting. This means the logfile logs look like this:
```
Oct 27 16:02:50.305 INFO Lighthouse started, version: Lighthouse/v2.0.1-8edd9d4+, module: lighthouse:413
Oct 27 16:02:50.305 INFO Configured for network, name: prater, module: lighthouse:414
```
2021-11-30 03:25:32 +00:00
Age Manning
6625aa4afe Status'd Peer Not Found (#2761)
## Issue Addressed

Users are experiencing `Status'd peer not found` errors

## Proposed Changes

Although I cannot reproduce this error, this is only one connection state change that is not addressed in the peer manager (that I could see). The error occurs because the number of disconnected peers in the peerdb becomes out of sync with the actual number of disconnected peers. From what I can tell almost all possible connection state changes are handled, except for the case when a disconnected peer changes to be disconnecting. This can potentially happen at the peer connection limit, where a previously connected peer switches to disconnecting. 

This PR decrements the disconnected counter when this event occurs and from what I can tell, covers all possible disconnection state changes in the peer manager.
2021-11-28 22:46:17 +00:00
Divma
413b0b5b2b Correctly update range status when outdated chains are removed (#2827)
We were batch removing chains when purging, and then updating the status of the collection for each of those. This makes the range status be out of sync with the real status. This represented no harm to the global sync status, but I've changed it to comply with a correct debug assertion that I got triggered while doing some testing.
Also added tests and improved code quality as per @paulhauner 's suggestions.
2021-11-26 01:13:49 +00:00
Pawan Dhananjay
9eedb6b888 Allow additional subnet peers (#2823)
## Issue Addressed

N/A

## Proposed Changes

1. Don't disconnect peer from dht on connection limit errors
2. Bump up `PRIORITY_PEER_EXCESS` to allow for dialing upto 60 peers by default.



Co-authored-by: Diva M <divma@protonmail.com>
2021-11-25 21:27:08 +00:00
Michael Sproul
2c07a72980 Revert peer DB changes from #2724 (#2828)
## Proposed Changes

This reverts commit 53562010ec from PR #2724

Hopefully this will restore the reliability of the sync simulator.
2021-11-25 03:45:52 +00:00
Michael Sproul
3fb8162dcc Use published ssz/tree_hash (#2825)
## Proposed Changes

Switch over to the latest published versions of the crates in the SSZ/`tree_hash` family.

## Additional Info

The crates were published at the current head of `unstable`: 0b319d4926. All 5 crates listed in this PR were published via tags, e.g. https://github.com/sigp/lighthouse/releases/tag/tree-hash-v0.4.0
2021-11-24 00:28:30 +00:00
Mac L
c61fbf71a4 Ensure consistent log formatting (#2819)
## Issue Addressed

N/A

## Proposed Changes

Filter out certain ascii characters when logging to ensure proper log formatting.
2021-11-23 07:06:22 +00:00
Age Manning
0b319d4926 Inform dialing via the behaviour (#2814)
I had this change but it seems to have been lost in chaos of network upgrades.

The swarm dialing event seems to miss some cases where we dial via the behaviour. This causes an error to be logged as the peer manager doesn't know about some dialing events. 

This shifts the logic to the behaviour to inform the peer manager.
2021-11-19 04:42:33 +00:00
Divma
53562010ec Move peer db writes to eth2 libp2p (#2724)
## Issue Addressed
Part of a bigger effort to make the network globals read only. This moves all writes to the `PeerDB` to the `eth2_libp2p` crate. Limiting writes to the peer manager is a slightly more complicated issue for a next PR, to keep things reviewable.

## Proposed Changes
- Make the peers field in the globals a private field.
- Allow mutable access to the peers field to `eth2_libp2p` for now.
- Add a new network message to update the sync state.

Co-authored-by: Age Manning <Age@AgeManning.com>
2021-11-19 04:42:31 +00:00
Divma
31386277c3 Sync wrong dbg assertion (#2821)
## Issue Addressed

Running a beacon node I triggered a sync debug panic. And so finally the time to create tests for sync arrived. Fortunately, te bug was not in the sync algorithm itself but a wrong assertion

## Proposed Changes

- Split Range's impl from the BeaconChain via a trait. This is needed for testing. The TestingRig/Harness is way bigger than needed and does not provide the modification functionalities that are needed to test sync. I find this simpler, tho some could disagree.
- Add a regression test for sync that fails before the changes.
- Fix the wrong assertion.
2021-11-19 02:38:25 +00:00
Age Manning
e519af9012 Update Lighthouse Dependencies (#2818)
## Issue Addressed

Updates lighthouse dependencies to resolve audit issues in out-dated deps.
2021-11-18 05:08:42 +00:00
Pawan Dhananjay
e32c09bfda Fix decoding max length (#2816)
## Issue Addressed

N/A

## Proposed Changes

Fix encoder max length to the correct value (`MAX_RPC_SIZE`).
2021-11-16 22:23:39 +00:00
Age Manning
a43a2448b7 Investigate and correct RPC Response Timeouts (#2804)
RPC Responses are for some reason not removing their timeout when they are completing. 

As an example:

```
Nov 09 01:18:20.256 DEBG Received BlocksByRange Request          step: 1, start_slot: 728465, count: 64, peer_id: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw
Nov 09 01:18:20.263 DEBG Received BlocksByRange Request          step: 1, start_slot: 728593, count: 64, peer_id: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw
Nov 09 01:18:20.483 DEBG BlocksByRange Response sent             returned: 63, requested: 64, current_slot: 2466389, start_slot: 728465, msg: Failed to return all requested blocks, peer: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw
Nov 09 01:18:20.500 DEBG BlocksByRange Response sent             returned: 64, requested: 64, current_slot: 2466389, start_slot: 728593, peer: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw
Nov 09 01:18:21.068 DEBG Received BlocksByRange Request          step: 1, start_slot: 728529, count: 64, peer_id: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw
Nov 09 01:18:21.272 DEBG BlocksByRange Response sent             returned: 63, requested: 64, current_slot: 2466389, start_slot: 728529, msg: Failed to return all requested blocks, peer: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw
Nov 09 01:18:23.434 DEBG Received BlocksByRange Request          step: 1, start_slot: 728657, count: 64, peer_id: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw
Nov 09 01:18:23.665 DEBG BlocksByRange Response sent             returned: 64, requested: 64, current_slot: 2466390, start_slot: 728657, peer: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw
Nov 09 01:18:25.851 DEBG Received BlocksByRange Request          step: 1, start_slot: 728337, count: 64, peer_id: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw
Nov 09 01:18:25.851 DEBG Received BlocksByRange Request          step: 1, start_slot: 728401, count: 64, peer_id: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw
Nov 09 01:18:26.094 DEBG BlocksByRange Response sent             returned: 62, requested: 64, current_slot: 2466390, start_slot: 728401, msg: Failed to return all requested blocks, peer: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw
Nov 09 01:18:26.100 DEBG BlocksByRange Response sent             returned: 63, requested: 64, current_slot: 2466390, start_slot: 728337, msg: Failed to return all requested blocks, peer: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw
Nov 09 01:18:31.070 DEBG RPC Error                               direction: Incoming, score: 0, peer_id: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw, client: Prysm: version: a80b1c252a9b4773493b41999769bf3134ac373f, os_version: unknown, err: Stream Timeout, protocol: beacon_blocks_by_range, service: libp2p
Nov 09 01:18:31.070 WARN Timed out to a peer's request. Likely insufficient resources, reduce peer count, service: libp2p
Nov 09 01:18:31.085 DEBG RPC Error                               direction: Incoming, score: 0, peer_id: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw, client: Prysm: version: a80b1c252a9b4773493b41999769bf3134ac373f, os_version: unknown, err: Stream Timeout, protocol: beacon_blocks_by_range, service: libp2p
Nov 09 01:18:31.085 WARN Timed out to a peer's request. Likely insufficient resources, reduce peer count, service: libp2p
Nov 09 01:18:31.459 DEBG RPC Error                               direction: Incoming, score: 0, peer_id: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw, client: Prysm: version: a80b1c252a9b4773493b41999769bf3134ac373f, os_version: unknown, err: Stream Timeout, protocol: beacon_blocks_by_range, service: libp2p
Nov 09 01:18:31.459 WARN Timed out to a peer's request. Likely insufficient resources, reduce peer count, service: libp2p
Nov 09 01:18:34.129 DEBG RPC Error                               direction: Incoming, score: 0, peer_id: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw, client: Prysm: version: a80b1c252a9b4773493b41999769bf3134ac373f, os_version: unknown, err: Stream Timeout, protocol: beacon_blocks_by_range, service: libp2p
Nov 09 01:18:34.130 WARN Timed out to a peer's request. Likely insufficient resources, reduce peer count, service: libp2p
Nov 09 01:18:35.686 DEBG Peer Manager disconnecting peer         reason: Too many peers, peer_id: 16Uiu2HAmEmBURejquBUMgKAqxViNoPnSptTWLA2CfgSPnnKENBNw, service: libp2p
```

This PR is to investigate and correct the issue. 

~~My current thoughts are that for some reason we are not closing the streams correctly, or fast enough, or the executor is not registering the closes and waking up.~~ - Pretty sure this is not the case, see message below for a more accurate reason.

~~I've currently added a timeout to stream closures in an attempt to force streams to close and the future to always complete.~~ I removed this
2021-11-16 03:42:25 +00:00
Pawan Dhananjay
58b04acf28 Fix simulator issues (#2802)
## Issue Addressed

Resolves #2775 

## Proposed Changes

1. Reduces the target_peers to stop recursive discovery requests.
2. Changes the eth1 `auto_update_interval` config parameter to be equal to the eth1 block time in the simulator. Without this, the eth1 `latest_cached_block` wouldn't be updated for some slots around `eth1_voting_period` boundaries which would lead to the eth1 cache falling out of sync. This is what caused the `Syncing eth1 block cache` and `No valid eth1_data votes` logs.
2021-11-15 06:38:29 +00:00
Pawan Dhananjay
3919046731 Add a section on using infura as the checkpoint sync provider (#2797)
## Issue Addressed

N/A

## Proposed Changes

From discord, it seems like users are a bit unclear on how to run checkpoint sync if they don't have an existing synced beacon node. Adds a note on how to use infura for the checkpoint sync feature.
2021-11-10 23:43:12 +00:00
Paul Hauner
931daa40d7 Add fork choice EF tests (#2737)
## Issue Addressed

Resolves #2545

## Proposed Changes

Adds the long-overdue EF tests for fork choice. Although we had pretty good coverage via other implementations that closely followed our approach, it is nonetheless important for us to implement these tests too.

During testing I found that we were using a hard-coded `SAFE_SLOTS_TO_UPDATE_JUSTIFIED` value rather than one from the `ChainSpec`. This caused a failure during a minimal preset test. This doesn't represent a risk to mainnet or testnets, since the hard-coded value matched the mainnet preset.

## Failing Cases

There is one failing case which is presently marked as `SkippedKnownFailure`:

```
case 4 ("new_finalized_slot_is_justified_checkpoint_ancestor") from /home/paul/development/lighthouse/testing/ef_tests/consensus-spec-tests/tests/minimal/phase0/fork_choice/on_block/pyspec_tests/new_finalized_slot_is_justified_checkpoint_ancestor failed with NotEqual:
head check failed: Got Head { slot: Slot(40), root: 0x9183dbaed4191a862bd307d476e687277fc08469fc38618699863333487703e7 } | Expected Head { slot: Slot(24), root: 0x105b49b51bf7103c182aa58860b039550a89c05a4675992e2af703bd02c84570 }
```

This failure is due to #2741. It's not a particularly high priority issue at the moment, so we fix it after merging this PR.
2021-11-08 07:29:04 +00:00
mooori
d01fe02824 Add regression tests for boot_node (#2749)
## Issue Addressed
Resolves #2602

## Proposed Changes

*Note: For a review it might help to look at the individual commits.*

### `boot_node`
Add support for the flags `dump-config` and `immediate-shutdown`. For `immediate-shutdown` the actual behavior could be described as `dump-config-and-exit`.

Both flags are handled in `boot_node::main`, which appears to be the simplest approach.

### `boot_node` regression tests
Added in `lighthouse/tests/boot_node.rs`.

### `CommandLineTestExec`
Factors out boilerplate related to CLI tests. It's used in the regression tests for `boot_node`, `beacon_node` and `validator_client`.

## Open TODO
Add tests for `boot_node` flags `enable-enr-auto-update` and `disable-packet-filter`. They end up in [`Discv5Config`](9ed2cba6bc/boot_node/src/config.rs (L29)), which doesn't support serde (de)serialization.

I haven't found a workaround - guidance would be appreciated.
2021-11-08 01:37:58 +00:00
Divma
fbafe416d1 Move the peer manager to be a behaviour (#2773)
This simply moves some functions that were "swarm notifications" to a network behaviour implementation.

Notes
------
- We could disconnect from the peer manager but we would lose the rpc shutdown message
- We still notify from the swarm since this is the most reliable way to get some events. Ugly but best for now
- Events need to be pushed with "add event" to wake the waker

Co-authored-by: Divma <26765164+divagant-martian@users.noreply.github.com>
2021-11-08 00:01:10 +00:00
Michael Sproul
df02639b71 De-duplicate attestations in the slasher (#2767)
## Issue Addressed

Closes https://github.com/sigp/lighthouse/issues/2112
Closes https://github.com/sigp/lighthouse/issues/1861

## Proposed Changes

Collect attestations by validator index in the slasher, and use the magic of reference counting to automatically discard redundant attestations. This results in us storing only 1-2% of the attestations observed when subscribed to all subnets, which carries over to a 50-100x reduction in data stored 🎉 

## Additional Info

There's some nuance to the configuration of the `slot-offset`. It has a profound effect on the effictiveness of de-duplication, see the docs added to the book for an explanation: 5442e695e5/book/src/slasher.md (slot-offset)
2021-11-08 00:01:09 +00:00
Rémy Roy
fadb8b2b2b Add minimum supported version for checkpoint sync in book (#2779)
## Issue Addressed

No specific issue. Just some improvement in the documentation provided by the book.

## Proposed Changes

Add minimum supported version for checkpoint sync in book to make sure users who want to use this feature know they need to be using at least version 2.0.0.
2021-11-05 06:35:52 +00:00
Pascal Bach
2ed6775dcf Add cargo vendor test (#2076)
## Issue Addressed

This is related to #1926 and #1712.

## Proposed Changes

This PR adds a test that make sure that the used dependencies can be vendored.

Being able to vendor the dependencies is important for archival and repdroducibility purpose.
It's also required to package lighthouse for some Linux distributions. Specifically [NixOS](https://nixos.org/) and [Yocto](https://www.yoctoproject.org/).

## Additional Info

This PR only adds the test, it doesn't clean up the dependencies yet. That's why it is in draft.
2021-11-05 04:42:12 +00:00
Divma
a683e0296a Peer manager cfg (#2766)
## Issue Addressed
I've done this change in a couple of WIPs already so I might as well submit it on its own. This changes no functionality but reduces coupling in a 0.0001%. It also helps new people who need to work in the peer manager to better understand what it actually needs from the outside

## Proposed Changes

Add a config to the peer manager
2021-11-03 23:44:44 +00:00
Divma
7502970a7d Do not compute metrics in the network service if the cli flag is not set (#2765)
## Issue Addressed

The computation of metrics in the network service can be expensive. This disables the computation unless the cli flag `metrics` is set.

## Additional Info
Metrics in other parts of the network are still updated, since most are simple metrics and checking if metrics are enabled each time each metric is updated doesn't seem like a gain.
2021-11-03 00:06:03 +00:00
realbigsean
c4ad0e3fb3 Ensure dependent root consistency in head events (#2753)
## Issue Addressed

@paulhauner noticed that when we send head events, we use the block root from `new_head` in `fork_choice_internal`, but calculate `dependent_root` and `previous_dependent_root` using the `canonical_head`. This is normally fine because `new_head` updates the `canonical_head` in `fork_choice_internal`, but it's possible we have a reorg updating `canonical_head` before our head events are sent. So this PR ensures `dependent_root` and `previous_dependent_root` are always derived from the state associated with `new_head`.



Co-authored-by: realbigsean <seananderson33@gmail.com>
2021-11-02 02:26:32 +00:00
Mac L
80627b428b Fix linting error on Windows (#2759)
While testing some code on Windows, I ran into a failure when using `clippy` via (`make lint`):
```
error: this expression borrows a reference (`&str`) that is immediately dereferenced by the compiler
   --> common/filesystem/src/lib.rs:105:43
    |
105 |         let mut acl = ACL::from_file_path(&path_str, false).map_err(Error::UnableToRetrieveACL)?;
    |                                           ^^^^^^^^^ help: change this to: `path_str`
    |
    = note: `-D clippy::needless-borrow` implied by `-D warnings`
    = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#needless_borrow
    
error: could not compile `filesystem` due to previous error
```

## Proposed Changes

Remove the unnecessary borrow as suggested.

## Additional Info

Since we are only running `clippy` in CI on Ubuntu, I believe we don't have any way (in CI) to detect these Windows specific lint errors (either from new code, or from linting changes from new Rust versions. 
This is because code marked as `#[cfg(windows)]` is not checked on `unix` systems and vice versa.

I'm conscious that our CI runs are already taking a long time, and that adding a new Windows `clippy` run would add a non-negligible amount of time to the runs (not sure if this topic has already been discussed), but it something to be aware of.

## Extra Note
I don't think this is the case, but it might be worth someone else running `make lint` on their Windows machines to eliminate the possibility that this is an error specific to my setup.
2021-11-01 07:44:42 +00:00
Pawan Dhananjay
4499adc7fd Check proposer index during block production (#2740)
## Issue Addressed

Resolves #2612 

## Proposed Changes

Implements both the checks mentioned in the original issue. 
1. Verifies the `randao_reveal` in the beacon node
2. Cross checks the proposer index after getting back the block from the beacon node.

## Additional info
The block production time increases by ~10x because of the signature verification on the beacon node (based on the `beacon_block_production_process_seconds` metric) when running on a local testnet.
2021-11-01 07:44:40 +00:00
Michael Sproul
ffb04e1a9e Add op pool metrics for attestations (#2758)
## Proposed Changes

Add several metrics for the number of attestations in the op pool. These give us a way to observe the number of valid, non-trivial attestations during block packing rather than just the size of the entire op pool.
2021-11-01 05:52:31 +00:00
Michael Sproul
e8a557fdd8 Update prometheus to v0.13.0 (#2757)
## Issue Addressed

Closes #2485
2021-11-01 05:52:30 +00:00
Divma
e2c0650d16 Relax late sync committee penalty (#2752)
## Issue Addressed

Getting too many peers kicked due to slightly late sync committee messages as tested on.. under-performant hardware.

## Proposed Changes

Only penalize if the message is more than one slot late. Still ignore the message-

Co-authored-by: Divma <26765164+divagant-martian@users.noreply.github.com>
2021-10-31 22:30:19 +00:00
Age Manning
1790010260 Upgrade to latest libp2p (#2605)
This is a pre-cursor to the next libp2p upgrade. 

It is currently being used for staging a number of PR upgrades which are contingent on the latest libp2p.
2021-10-29 01:59:29 +00:00
ethDreamer
2c4413454a Fixed Gossip Topics on Fork Boundary (#2619)
## Issue Addressed

The [p2p-interface section of the `altair` spec](https://github.com/ethereum/consensus-specs/blob/dev/specs/altair/p2p-interface.md#transitioning-the-gossip) says you should subscribe to the topics for a fork "In advance of the fork" and unsubscribe from old topics `2 Epochs` after the new fork is activated. We've chosen to subscribe to new fork topics `2 slots` before the fork is initiated.

This function is supposed to return the required fork digests at any given time but as it was currently written, it doesn't return the fork digest for a previous fork if you've switched to the current fork less than 2 epoch's ago. Also this function required modification for every new fork we add.

## Proposed Changes

Make this function fork-agnostic and correctly handle the previous fork topic digests when you've only just switched to the new fork.
2021-10-29 00:05:27 +00:00
Pawan Dhananjay
88063398f6 Prevent double import of blocks (#2647)
## Issue Addressed

Resolves #2611 

## Proposed Changes

Adds a duplicate block root cache to the `BeaconProcessor`. Adds the block root to the cache before calling `process_gossip_block` and `process_rpc_block`. Since `process_rpc_block` is called only for single block lookups, we don't have to worry about batched block imports.

The block is imported from the source(gossip/rpc) that arrives first. The block that arrives second is not imported to avoid the db access issue.
There are 2 cases:
1. Block that arrives second is from rpc: In this case, we return an optimistic `BlockError::BlockIsAlreadyKnown` to sync.
2. Block that arrives second is from gossip: In this case, we only do gossip verification and forwarding but don't import the block into the the beacon chain.

## Additional info
Splits up `process_gossip_block` function to `process_gossip_unverified_block` and `process_gossip_verified_block`.
2021-10-28 03:36:14 +00:00