See related issue: https://github.com/ethpandaops/dora/pull/713
When LH emits a `head` event the block isn't written to disk yet. Some upstream consumers may expect that after a `head` event that the block should be queryable via the beacon api. This PR falls back to fetching the block from the early attester cache if it wasn't found in the store. This should ensure that a block is always queryable immediately after a `head` event is emitted.
Additionally I noticed that when serving columns we always default to using the store. We already have `get_data_columns_checking_all_caches ` which tries the da cache, then the store and finally the early attester cache.
Co-Authored-By: Eitan Seri-Levi <eserilev@ucsc.edu>
Co-Authored-By: Michael Sproul <michael@sigmaprime.io>
Breakout from:
- https://github.com/sigp/lighthouse/pull/9295
We currently do not handle the verification of payload attestations on non-canonical side chains, we always attempt to use the head. The included regression test demonstrates this, and there is _also_ a fork choice compliance test in #9295 that triggers it.
This PR is a bit opinionated, but I'll explain my judgements:
- We need a way to get the PTC for an arbitrary slot from an arbitrary state. This involves potential state advances, database lookups, etc. There is some fiddly logic required to check that states are in range/etc.
- We _already have_ a cache with the exact same lifecycle as the PTCs, namely the attester shuffling cache. Therefore, we can de-duplicate a lot of the complexity by storing the PTCs for a given epoch (and decision block) in this cache.
The other opinionated change is in the tests. The previous tests were set up kind of nicely to avoid instantiating a `BeaconChainHarness`. However they were not using mocking, which made testing the non-canonical chain case kind of infeasible. To remedy this, I've changed them to just use a beacon chain harness and create two chains using its relatively easy to use methods for doing this. The running time of the tests goes from something like 2.6s for 8 tests to 3.3s for 9 tests, which is only an increase of 0.04s/test. Negligible. Another plus to using the `BeaconChainHarness` is that it avoids a bunch of the cruft to create synthetic non-mocked beacon chain bits.
At the same time, I've made some attempt to improve modularity (and fit with the `GossipVerificationContext`) by pulling out the guts of `with_committee_cache` into a new function (`with_cached_shuffling`) that clearly shows its dependency surface.
Co-Authored-By: Michael Sproul <michael@sigmaprime.io>
Co-Authored-By: dapplion <35266934+dapplion@users.noreply.github.com>
Use a macro to remove the repetitive fork variant boilerplate in `signed_beacon_block.rs` when implementing `into_full_block` for the various `SignedBeaconBlock` variants
Co-Authored-By: Mac L <mjladson@pm.me>
Add a new feature flag to `lighthouse` which adds jemalloc profiling support.
We could manually add this during memory profiling but it is a nice QoL to have this built-in imo
Co-Authored-By: Mac L <mjladson@pm.me>
Peers that advertise that they have imported a block may not have the columns for that slot available post-Gloas. Ensure that we dont penalize them.
Co-Authored-By: Eitan Seri-Levi <eserilev@ucsc.edu>
During custody backfill sync if a peer fails to serve columns for a batch don't penalize them more than once per batch
Co-Authored-By: Eitan Seri-Levi <eserilev@ucsc.edu>
Using `state.fork` is a bit sketchy at the fork boundary. It's safer to just use the payload attestations slot
Co-Authored-By: Eitan Seri-Levi <eserilev@ucsc.edu>
Forgot to add `ENGINE_GET_BLOBS_V3` to `LIGHTHOUSE_CAPABILITIES`.
Add `ENGINE_GET_BLOBS_V3` to `LIGHTHOUSE_CAPABILITIES`.
Co-Authored-By: Daniel Knopik <daniel@dknopik.de>
CachedPTCs::try_from_state now returns Result<Option<Self>, _> and
internalises the boundary rule (pre-Gloas state, Gloas shuffling epoch
=> Ok(None)). Callers (block import priming, state advance timer,
with_cached_shuffling miss path) just skip insertion on None instead
of duplicating the guard. The unit test exercises the three boundary
cases against a pre-Gloas state.
- Remove unused `BeaconChainError::MissingPtcForGloasShuffling` variant
(no producers remained after the earlier cleanup).
- Drop the `Result<(), BeaconChainError>` return type from
`ShufflingCache::insert_committee_cache`; both match arms are
infallible. Update callers in `beacon_chain.rs`, `state_advance_timer.rs`,
`shuffling_cache.rs` and the unit tests accordingly.
- Trim stale "Replace the committee if it's not present" comment in
`insert_committee_cache`; the Committee arm is now a no-op so only
the `Promise(_) | None` whimsy line remains.
#8314 left a few ugly potentially panicking location behind - all of them believed to be unreachable, but this PR fixes them regardless for good hygiene.
Update to `ethereum_ssz 0.10.4` for two new helpers: `not_inplace` and `clone_zeroed`.
Remove remaining `expect` and `todo!` in favour of these helpers and one new fallible (but practically infallible) method.
Co-Authored-By: Daniel Knopik <daniel@dknopik.de>
- PR https://github.com/sigp/lighthouse/pull/9305 wants to store PTCs in the committee cache.
BUT the http API route wants to use the committee cache and insert historical committees (i.e. given state at epoch 1000, compute and store the committee for epoch 900).
If we want a single cache to serve both use cases we need to:
- Have entries in the committee cache that have no PTC: Makes reading PTCs from the cache not deterministic
- Compute historical PTC: A bunch of complicated code that's useless
Instead we can add a separate cache for the API, very simple one, that caches committees only. And have the one in the beacon chain compute and cache PTCs always.
### Performance impact
Slightly additional memory cost for users of the `beacon/states/committees` route. Caching is almost equivalent, except for queries of recent committees that may already exist in the beacon chain's committee cache.
### AI disclousure
This PR was written by hand 90%. Claude fixed some warp type issues
Co-Authored-By: dapplion <35266934+dapplion@users.noreply.github.com>
While working on this code in another branch I noticed we had this messy, complicated and incorrect code about SPRP (slots-per-restore-point), which is no longer a relevant concept since the introduction of hot state diffs.
In the name of simplicity, I've removed any kind of hinting here in favour of a simple out of bounds error. The benefit of adding complex hinting code (which is not tested) to such a function is not worth it IMO. Users will work it out (or ask) if we just tell them their request is out of bounds.
Co-Authored-By: Michael Sproul <michael@sigmaprime.io>
In Gloas, beacon blocks are imported into fork choice immediately - the payload envelope and data columns arrive
separately. KZG commitments moved from the column sidecar into the execution payload bid, so the existing
`DataAvailabilityChecker` (which assumes block and data are coupled) can't be used for Gloas.
* Introduced `PendingPayloadCache` to keep track of payload and data columns per block root.
* Added gossip column verification
* Added support for Gloas data column reconstruction
* Payload envelope verification simplified: removed `MaybeAvailableEnvelope`, `ExecutedEnvelope`, `EnvelopeImportData`
Not yet implemented (tracked with TODOs):
- Proper lookup sync for Gloas columns arriving before blocks
- Partial column merging for Gloas
- Moving `load_gloas_payload_bid` disk reads off the async runtime
- Backfill/range sync for Gloas
Based on @eserilev's PR and work in progress. See also #9202 for verification.
Co-Authored-By: Eitan Seri-Levi <eserilev@ucsc.edu>
Co-Authored-By: Eitan Seri- Levi <eserilev@gmail.com>
Co-Authored-By: Daniel Knopik <daniel@dknopik.de>
Co-Authored-By: Daniel Knopik <107140945+dknopik@users.noreply.github.com>
Co-Authored-By: dapplion <35266934+dapplion@users.noreply.github.com>
Co-Authored-By: Jimmy Chen <jchen.tc@gmail.com>
Using the `ignore-ws-check` doesn't actually let you start up a node thats outside the weak subjectivity period
Co-Authored-By: Eitan Seri-Levi <eserilev@ucsc.edu>
Test helpers `add_attested_block_at_slot` and `add_attested_blocks_at_slot` accepted `state_root` argument which was computed before applying the block.
Co-Authored-By: hopinheimer <knmanas6@gmail.com>
