## Issue Addressed
NA
## Proposed Changes
Rather than spawning new tasks on the tokio executor to process each HTTP API request, send the tasks to the `BeaconProcessor`. This achieves:
1. Places a bound on how many concurrent requests are being served (i.e., how many we are actually trying to compute at one time).
1. Places a bound on how many requests can be awaiting a response at one time (i.e., starts dropping requests when we have too many queued).
1. Allows the BN prioritise HTTP requests with respect to messages coming from the P2P network (i.e., proiritise importing gossip blocks rather than serving API requests).
Presently there are two levels of priorities:
- `Priority::P0`
- The beacon processor will prioritise these above everything other than importing new blocks.
- Roughly all validator-sensitive endpoints.
- `Priority::P1`
- The beacon processor will prioritise practically all other P2P messages over these, except for historical backfill things.
- Everything that's not `Priority::P0`
The `--http-enable-beacon-processor false` flag can be supplied to revert back to the old behaviour of spawning new `tokio` tasks for each request:
```
--http-enable-beacon-processor <BOOLEAN>
The beacon processor is a scheduler which provides quality-of-service and DoS protection. When set to
"true", HTTP API requests will queued and scheduled alongside other tasks. When set to "false", HTTP API
responses will be executed immediately. [default: true]
```
## New CLI Flags
I added some other new CLI flags:
```
--beacon-processor-aggregate-batch-size <INTEGER>
Specifies the number of gossip aggregate attestations in a signature verification batch. Higher values may
reduce CPU usage in a healthy network while lower values may increase CPU usage in an unhealthy or hostile
network. [default: 64]
--beacon-processor-attestation-batch-size <INTEGER>
Specifies the number of gossip attestations in a signature verification batch. Higher values may reduce CPU
usage in a healthy network whilst lower values may increase CPU usage in an unhealthy or hostile network.
[default: 64]
--beacon-processor-max-workers <INTEGER>
Specifies the maximum concurrent tasks for the task scheduler. Increasing this value may increase resource
consumption. Reducing the value may result in decreased resource usage and diminished performance. The
default value is the number of logical CPU cores on the host.
--beacon-processor-reprocess-queue-len <INTEGER>
Specifies the length of the queue for messages requiring delayed processing. Higher values may prevent
messages from being dropped while lower values may help protect the node from becoming overwhelmed.
[default: 12288]
```
I needed to add the max-workers flag since the "simulator" flavor tests started failing with HTTP timeouts on the test assertions. I believe they were failing because the Github runners only have 2 cores and there just weren't enough workers available to process our requests in time. I added the other flags since they seem fun to fiddle with.
## Additional Info
I bumped the timeouts on the "simulator" flavor test from 4s to 8s. The prioritisation of consensus messages seems to be causing slower responses, I guess this is what we signed up for 🤷
The `validator/register` validator has some special handling because the relays have a bad habit of timing out on these calls. It seems like a waste of a `BeaconProcessor` worker to just wait for the builder API HTTP response, so we spawn a new `tokio` task to wait for a builder response.
I've added an optimisation for the `GET beacon/states/{state_id}/validators/{validator_id}` endpoint in [efbabe3](efbabe3252). That's the endpoint the VC uses to resolve pubkeys to validator indices, and it's the endpoint that was causing us grief. Perhaps I should move that into a new PR, not sure.
## Issue Addressed
Fix an issue observed by `@zlan` on Discord where Lighthouse would sometimes return this error when looking up states via the API:
> {"code":500,"message":"UNHANDLED_ERROR: ForkChoiceError(MissingProtoArrayBlock(0xc9cf1495421b6ef3215d82253b388d77321176a1dcef0db0e71a0cd0ffc8cdb7))","stacktraces":[]}
## Proposed Changes
The error stems from a faulty assumption in the HTTP API logic: that any state in the hot database must have its block in fork choice. This isn't true because the state's hot database may update much less frequently than the fork choice store, e.g. if reconstructing states (where freezer migration pauses), or if the freezer migration runs slowly. There could also be a race between loading the hot state and checking fork choice, e.g. even if the finalization migration of DB+fork choice were atomic, the update could happen between the 1st and 2nd calls.
To address this I've changed the HTTP API logic to use the finalized block's execution status as a fallback where it is safe to do so. In the case where a block is non-canonical and prior to finalization (permanently orphaned) we default `execution_optimistic` to `true`.
## Additional Info
I've also added a new CLI flag to reduce the frequency of the finalization migration as this is useful for several purposes:
- Spacing out database writes (less frequent, larger batches)
- Keeping a limited chain history with high availability, e.g. the last month in the hot database.
This new flag made it _substantially_ easier to test this change. It was extracted from `tree-states` (where it's called `--db-migration-period`), which is why this PR also carries the `tree-states` label.
## Issue Addressed
#4118
## Proposed Changes
This PR introduces a "progressive balances" cache on the `BeaconState`, which keeps track of the accumulated target attestation balance for the current & previous epochs. The cached values are utilised by fork choice to calculate unrealized justification and finalization (instead of converting epoch participation arrays to balances for each block we receive).
This optimization will be rolled out gradually to allow for more testing. A new `--progressive-balances disabled|checked|strict|fast` flag is introduced to support this:
- `checked`: enabled with checks against participation cache, and falls back to the existing epoch processing calculation if there is a total target attester balance mismatch. There is no performance gain from this as the participation cache still needs to be computed. **This is the default mode for now.**
- `strict`: enabled with checks against participation cache, returns error if there is a mismatch. **Used for testing only**.
- `fast`: enabled with no comparative checks and without computing the participation cache. This mode gives us the performance gains from the optimization. This is still experimental and not currently recommended for production usage, but will become the default mode in a future release.
- `disabled`: disable the usage of progressive cache, and use the existing method for FFG progression calculation. This mode may be useful if we find a bug and want to stop the frequent error logs.
### Tasks
- [x] Initial cache implementation in `BeaconState`
- [x] Perform checks in fork choice to compare the progressive balances cache against results from `ParticipationCache`
- [x] Add CLI flag, and disable the optimization by default
- [x] Testing on Goerli & Benchmarking
- [x] Move caching logic from state processing to the `ProgressiveBalancesCache` (see [this comment](https://github.com/sigp/lighthouse/pull/4362#discussion_r1230877001))
- [x] Add attesting balance metrics
Co-authored-by: Jimmy Chen <jimmy@sigmaprime.io>
## Proposed Changes
Remove `max-skip-slots` checks when processing blocks.
This was legacy code which was previously used in the Medalla testnet to sync to the correct fork.
With the addition of checkpoint sync which allows us to sync to any arbitrary fork, this is no longer a necessary feature, so it has been removed for simplicity.
## Additional Notes
The CLI flag and checks for attestation processing have been retained as it still may have uses in DoS protection.
## 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>
## 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
#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
## Issue Addressed
NA
## Proposed Changes
- Implements https://github.com/ethereum/consensus-specs/pull/3290/
- Bumps `ef-tests` to [v1.3.0-rc.4](https://github.com/ethereum/consensus-spec-tests/releases/tag/v1.3.0-rc.4).
The `CountRealizedFull` concept has been removed and the `--count-unrealized-full` and `--count-unrealized` BN flags now do nothing but log a `WARN` when used.
## Database Migration Debt
This PR removes the `best_justified_checkpoint` from fork choice. This field is persisted on-disk and the correct way to go about this would be to make a DB migration to remove the field. However, in this PR I've simply stubbed out the value with a junk value. I've taken this approach because if we're going to do a DB migration I'd love to remove the `Option`s around the justified and finalized checkpoints on `ProtoNode` whilst we're at it. Those options were added in #2822 which was included in Lighthouse v2.1.0. The options were only put there to handle the migration and they've been set to `Some` ever since v2.1.0. There's no reason to keep them as options anymore.
I started adding the DB migration to this branch but I started to feel like I was bloating this rather critical PR with nice-to-haves. I've kept the partially-complete migration [over in my repo](https://github.com/paulhauner/lighthouse/tree/fc-pr-18-migration) so we can pick it up after this PR is merged.
This PR enables the user to adjust the shuffling cache size.
This is useful for some HTTP API requests which require re-computing old shufflings. This PR currently optimizes the
beacon/states/{state_id}/committees HTTP API by first checking the cache before re-building shuffling.
If the shuffling is set to a non-default value, then the HTTP API request will also fill the cache when as it constructs new shufflings.
If the CLI flag is not present or the value is set to the default of 16 the default behaviour is observed.
Co-authored-by: Michael Sproul <michael@sigmaprime.io>
## Issue Addressed
Closes#3896Closes#3998Closes#3700
## Proposed Changes
- Optimise the calculation of withdrawals for payload attributes by avoiding state clones, avoiding unnecessary state advances and reading from the snapshot cache if possible.
- Use the execution layer's payload attributes cache to avoid re-calculating payload attributes. I actually implemented a new LRU cache just for withdrawals but it had the exact same key and most of the same data as the existing payload attributes cache, so I deleted it.
- Add a new SSE event that fires when payloadAttributes are calculated. This is useful for block builders, a la https://github.com/ethereum/beacon-APIs/issues/244.
- Add a new CLI flag `--always-prepare-payload` which forces payload attributes to be sent with every fcU regardless of connected proposers. This is intended for use by builders/relays.
For maximum effect, the flags I've been using to run Lighthouse in "payload builder mode" are:
```
--always-prepare-payload \
--prepare-payload-lookahead 12000 \
--suggested-fee-recipient 0x0000000000000000000000000000000000000000
```
The fee recipient is required so Lighthouse has something to pack in the payload attributes (it can be ignored by the builder). The lookahead causes fcU to be sent at the start of every slot rather than at 8s. As usual, fcU will also be sent after each change of head block. I think this combination is sufficient for builders to build on all viable heads. Often there will be two fcU (and two payload attributes) sent for the same slot: one sent at the start of the slot with the head from `n - 1` as the parent, and one sent after the block arrives with `n` as the parent.
Example usage of the new event stream:
```bash
curl -N "http://localhost:5052/eth/v1/events?topics=payload_attributes"
```
## Additional Info
- [x] Tests added by updating the proposer re-org tests. This has the benefit of testing the proposer re-org code paths with withdrawals too, confirming that the new changes don't interact poorly.
- [ ] Benchmarking with `blockdreamer` on devnet-7 showed promising results but I'm yet to do a comparison to `unstable`.
Co-authored-by: Michael Sproul <micsproul@gmail.com>
## Issue Addressed
Recent discussions with other client devs about optimistic sync have revealed a conceptual issue with the optimisation implemented in #3738. In designing that feature I failed to consider that the execution node checks the `blockHash` of the execution payload before responding with `SYNCING`, and that omitting this check entirely results in a degradation of the full node's validation. A node omitting the `blockHash` checks could be tricked by a supermajority of validators into following an invalid chain, something which is ordinarily impossible.
## Proposed Changes
I've added verification of the `payload.block_hash` in Lighthouse. In case of failure we log a warning and fall back to verifying the payload with the execution client.
I've used our existing dependency on `ethers_core` for RLP support, and a new dependency on Parity's `triehash` crate for the Merkle patricia trie. Although the `triehash` crate is currently unmaintained it seems like our best option at the moment (it is also used by Reth, and requires vastly less boilerplate than Parity's generic `trie-root` library).
Block hash verification is pretty quick, about 500us per block on my machine (mainnet).
The optimistic finalized sync feature can be disabled using `--disable-optimistic-finalized-sync` which forces full verification with the EL.
## Additional Info
This PR also introduces a new dependency on our [`metastruct`](https://github.com/sigp/metastruct) library, which was perfectly suited to the RLP serialization method. There will likely be changes as `metastruct` grows, but I think this is a good way to start dogfooding it.
I took inspiration from some Parity and Reth code while writing this, and have preserved the relevant license headers on the files containing code that was copied and modified.
## Proposed Changes
With proposer boosting implemented (#2822) we have an opportunity to re-org out late blocks.
This PR adds three flags to the BN to control this behaviour:
* `--disable-proposer-reorgs`: turn aggressive re-orging off (it's on by default).
* `--proposer-reorg-threshold N`: attempt to orphan blocks with less than N% of the committee vote. If this parameter isn't set then N defaults to 20% when the feature is enabled.
* `--proposer-reorg-epochs-since-finalization N`: only attempt to re-org late blocks when the number of epochs since finalization is less than or equal to N. The default is 2 epochs, meaning re-orgs will only be attempted when the chain is finalizing optimally.
For safety Lighthouse will only attempt a re-org under very specific conditions:
1. The block being proposed is 1 slot after the canonical head, and the canonical head is 1 slot after its parent. i.e. at slot `n + 1` rather than building on the block from slot `n` we build on the block from slot `n - 1`.
2. The current canonical head received less than N% of the committee vote. N should be set depending on the proposer boost fraction itself, the fraction of the network that is believed to be applying it, and the size of the largest entity that could be hoarding votes.
3. The current canonical head arrived after the attestation deadline from our perspective. This condition was only added to support suppression of forkchoiceUpdated messages, but makes intuitive sense.
4. The block is being proposed in the first 2 seconds of the slot. This gives it time to propagate and receive the proposer boost.
## Additional Info
For the initial idea and background, see: https://github.com/ethereum/consensus-specs/pull/2353#issuecomment-950238004
There is also a specification for this feature here: https://github.com/ethereum/consensus-specs/pull/3034
Co-authored-by: Michael Sproul <micsproul@gmail.com>
Co-authored-by: pawan <pawandhananjay@gmail.com>
## Issue Addressed
Partially addresses #3651
## Proposed Changes
Adds server-side support for light_client_bootstrap_v1 topic
## Additional Info
This PR, creates each time a bootstrap without using cache, I do not know how necessary a cache is in this case as this topic is not supposed to be called frequently and IMHO we can just prevent abuse by using the limiter, but let me know what you think or if there is any caveat to this, or if it is necessary only for the sake of good practice.
Co-authored-by: Pawan Dhananjay <pawandhananjay@gmail.com>
## Issue Addressed
Part of https://github.com/sigp/lighthouse/issues/3651.
## Proposed Changes
Add a flag for enabling the light client server, which should be checked before gossip/RPC traffic is processed (e.g. https://github.com/sigp/lighthouse/pull/3693, https://github.com/sigp/lighthouse/pull/3711). The flag is available at runtime from `beacon_chain.config.enable_light_client_server`.
Additionally, a new method `BeaconChain::with_mutable_state_for_block` is added which I envisage being used for computing light client updates. Unfortunately its performance will be quite poor on average because it will only run quickly with access to the tree hash cache. Each slot the tree hash cache is only available for a brief window of time between the head block being processed and the state advance at 9s in the slot. When the state advance happens the cache is moved and mutated to get ready for the next slot, which makes it no longer useful for merkle proofs related to the head block. Rather than spend more time trying to optimise this I think we should continue prototyping with this code, and I'll make sure `tree-states` is ready to ship before we enable the light client server in prod (cf. https://github.com/sigp/lighthouse/pull/3206).
## Additional Info
I also fixed a bug in the implementation of `BeaconState::compute_merkle_proof` whereby the tree hash cache was moved with `.take()` but never put back with `.restore()`.
## Issue Addressed
#3702
Which issue # does this PR address?
#3702
## Proposed Changes
Added checkpoint-sync-url-timeout flag to cli. Added timeout field to ClientGenesis::CheckpointSyncUrl to utilize timeout set
## Additional Info
Please provide any additional information. For example, future considerations
or information useful for reviewers.
Co-authored-by: GeemoCandama <104614073+GeemoCandama@users.noreply.github.com>
Co-authored-by: Michael Sproul <micsproul@gmail.com>
## Issue Addressed
Add a flag that can increase count unrealized strictness, defaults to false
## Proposed Changes
Please list or describe the changes introduced by this PR.
## Additional Info
Please provide any additional information. For example, future considerations
or information useful for reviewers.
Co-authored-by: realbigsean <seananderson33@gmail.com>
Co-authored-by: sean <seananderson33@gmail.com>
## 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.
## 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 52bb1840ae.
## 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.
## Proposed Changes
Update the invalid head tests so that they work with the current default fork choice configuration.
Thanks @realbigsean for fixing the persistence test and the EF tests.
Co-authored-by: realbigsean <sean@sigmaprime.io>
## Issue Addressed
https://github.com/sigp/lighthouse/issues/3091
Extends https://github.com/sigp/lighthouse/pull/3062, adding pre-bellatrix block support on blinded endpoints and allowing the normal proposal flow (local payload construction) on blinded endpoints. This resulted in better fallback logic because the VC will not have to switch endpoints on failure in the BN <> Builder API, the BN can just fallback immediately and without repeating block processing that it shouldn't need to. We can also keep VC fallback from the VC<>BN API's blinded endpoint to full endpoint.
## Proposed Changes
- Pre-bellatrix blocks on blinded endpoints
- Add a new `PayloadCache` to the execution layer
- Better fallback-from-builder logic
## Todos
- [x] Remove VC transition logic
- [x] Add logic to only enable builder flow after Merge transition finalization
- [x] Tests
- [x] Fix metrics
- [x] Rustdocs
Co-authored-by: Mac L <mjladson@pm.me>
Co-authored-by: realbigsean <sean@sigmaprime.io>
## Issue Addressed
Add a flag that optionally enables unrealized vote tracking. Would like to test out on testnets and benchmark differences in methods of vote tracking. This PR includes a DB schema upgrade to enable to new vote tracking style.
Co-authored-by: realbigsean <sean@sigmaprime.io>
Co-authored-by: Paul Hauner <paul@paulhauner.com>
Co-authored-by: sean <seananderson33@gmail.com>
Co-authored-by: Mac L <mjladson@pm.me>
## 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 😢 ). 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.
Fix max packet sizes
Fix max_payload_size function
Add merge block test
Fix max size calculation; fix up test
Clear comments
Add a payload_size_function
Use safe arith for payload calculation
Return an error if block too big in block production
Separate test to check if block is over limit
## Issue Addressed
Mitigates #1096
## Proposed Changes
Add a flag to the beacon node called `--disable-lock-timeouts` which allows opting out of lock timeouts.
The lock timeouts serve a dual purpose:
1. They prevent any single operation from hogging the lock for too long. When a timeout occurs it logs a nasty error which indicates that there's suboptimal lock use occurring, which we can then act on.
2. They allow deadlock detection. We're fairly sure there are no deadlocks left in Lighthouse anymore but the timeout locks offer a safeguard against that.
However, timeouts on locks are not without downsides:
They allow for the possibility of livelock, particularly on slower hardware. If lock timeouts keep failing spuriously the node can be prevented from making any progress, even if it would be able to make progress slowly without the timeout. One particularly concerning scenario which could occur would be if a DoS attack succeeded in slowing block signature verification times across the network, and all Lighthouse nodes got livelocked because they timed out repeatedly. This could also occur on just a subset of nodes (e.g. dual core VPSs or Raspberri Pis).
By making the behaviour runtime configurable this PR allows us to choose the behaviour we want depending on circumstance. I suspect that long term we could make the timeout-free approach the default (#2381 moves in this direction) and just enable the timeouts on our testnet nodes for debugging purposes. This PR conservatively leaves the default as-is so we can gain some more experience before switching the default.
## Issue Addressed
Closes#1891Closes#1784
## Proposed Changes
Implement checkpoint sync for Lighthouse, enabling it to start from a weak subjectivity checkpoint.
## Additional Info
- [x] Return unavailable status for out-of-range blocks requested by peers (#2561)
- [x] Implement sync daemon for fetching historical blocks (#2561)
- [x] Verify chain hashes (either in `historical_blocks.rs` or the calling module)
- [x] Consistency check for initial block + state
- [x] Fetch the initial state and block from a beacon node HTTP endpoint
- [x] Don't crash fetching beacon states by slot from the API
- [x] Background service for state reconstruction, triggered by CLI flag or API call.
Considered out of scope for this PR:
- Drop the requirement to provide the `--checkpoint-block` (this would require some pretty heavy refactoring of block verification)
Co-authored-by: Diva M <divma@protonmail.com>
## Issue Addressed
Closes#800Closes#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.
## Issue Addressed
Solution 2 proposed here: https://github.com/sigp/lighthouse/issues/1435#issuecomment-692317639
## Proposed Changes
- Adds an optional `--wss-checkpoint` flag that takes a string `root:epoch`
- Verify that the given checkpoint exists in the chain, or that the the chain syncs through this checkpoint. If not, shutdown and prompt the user to purge state before restarting.
## Additional Info
Co-authored-by: Paul Hauner <paul@paulhauner.com>
## Issue Addressed
NA
## Proposed Changes
Sets the default max skips to 700 so that it can cover the 693 slot skip from `80894 - 80201`.
## Additional Info
NA
## Proposed Changes
To mitigate the impact of minority forks on RAM and disk usage, this change rejects blocks whose parent lies more than 320 slots (10 epochs, ~1 hour) in the past. The behaviour is configurable via `lighthouse bn --max-skip-slots N`, and can be turned off entirely using `--max-skip-slots none`.
Co-authored-by: Paul Hauner <paul@paulhauner.com>
