Use block headers in cache handleHead, full blocks are best-effort now#397
Merged
Use block headers in cache handleHead, full blocks are best-effort now#397
Conversation
AntiD2ta
changed the title
AntiD2ta
force-pushed
the
fix/cache-block-header
branch
from
c160486 to
228ac27
Compare
Sync committee validators can occupy multiple positions in the same subcommittee, producing duplicate (account, root) pairs that Dirk rejects as a batch. Deduplicate before signing and scatter the signatures back to all original positions. Fixes attestantio/attestant#630
When a validator is assigned to multiple sync committee subnets, Vouch sends different signing roots (one per subnet) for the same account in a single SignGenericMulti call. Dirk's RunRules rejects any batch containing the same pubkey more than once, causing all signing to fail. After the existing (account, root) deduplication, split remaining entries into N batches where each batch contains at most one entry per account. The fast path (all accounts unique) is preserved with zero overhead. Each batch is sent as a separate SignGenericMulti call and signatures are mapped back to original positions.
Extract multi-signer batch-splitting logic into standalone signMulti function. Rename uniqueAccounts/uniqueRoots/uniqueData/uniqueSigs to uniquePairAccounts/uniquePairRoots/uniquePairData/uniquePairSigs to clarify that uniqueness is at the (account, root) pair level. Wrap SignGenericMulti errors consistently with the rest of the package. Add test cases for error propagation when a batch fails mid-signing.
Cover the scenario where all (account, root) pairs are identical, verifying dedup reduces them to a single unique entry and the numBatches == 1 fast path in signMulti is exercised.
Beacon nodes like Lighthouse prune block bodies but retain headers. The cache's handleHead was fetching full blocks via /eth/v2/beacon/blocks/<root>, which returns 404 for pruned blocks, causing persistent error-level logs and stale execution chain head data. Change handleHead and startup to fetch BeaconBlockHeader first (always available), caching block root to slot mapping. Full block fetch for execution payload data becomes best-effort — failure is logged at debug level since execution data is only used for builder bid gas limit verification and is not critical for core validator duties. Resolves #393.
Use data.Slot from the HeadEvent for the root-to-slot cache mapping instead of fetching headers via HTTP. This is consistent with handleBlock which also uses event data directly, and eliminates a redundant HTTP request per slot. Simplify tests to verify slot is always cached from the event regardless of whether the full block fetch succeeds or fails.
All callers already handle the returned error at their own appropriate log level. The strategy's internal Warn on timeout is redundant and produces noise when the strategy is used best-effort (e.g. cache handleHead on Lighthouse/Erigon where block bodies are pruned). This aligns with line 81 in the same function where individual provider failures already log at Debug.
…egies Move clientMonitor.ClientOperation before error type checking, matching the pattern in all other first strategies. Previously, 404 and 503 responses bypassed monitoring entirely, making pruned-block beacon nodes invisible in Prometheus metrics.
AntiD2ta
force-pushed
the
fix/cache-block-header
branch
from
228ac27 to
47f1103
Compare
AntiD2ta
added a commit
that referenced
this pull request
Apr 17, 2026
Bez625
requested changes
Apr 21, 2026
Bez625
approved these changes
Apr 22, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Problem
Vouch's cache
handleHeadfetches full signed beacon blocks via/eth/v2/beacon/blocks/<root>on every head event. Lighthouse prunes block bodies by default, returning 404 for this endpoint. Thesignedbeaconblock/firststrategy silently ignores 404s but never sends a response, causing a timeout and persistent error-level logs on every slot. The execution data extracted from the full block (gas limit, chain head) is only used for builder bid verification and is not critical for core validator duties.Solution
Use
data.Slotfrom theHeadEventdirectly for the root-to-slot cache (consistent withhandleBlock), avoiding an HTTP round-trip entirely. Attempt the full block fetch as best-effort for execution payload data, logging failures at debug level instead of error. The startup path inservice.go:Newuses theBeaconBlockHeaderendpoint since no event data is available at that point.Resolves #393.
Dependends on #394 which depends on on #396 for passing the Trivy scan.