I found myself having to update this code for Gloas, and figured we may as well delete it seeing as it doesn't work.
See:
- https://github.com/sigp/lighthouse/issues/4198
Delete all `fork_revert` logic and the accompanying test.
Co-Authored-By: Michael Sproul <michael@sigmaprime.io>
Which issue # does this PR address?
None
All of these are performing a check, and adding a batch, or creating a new lookup, or a new query, etc..
Hence all of these limits would be off by one.
Example:
```rust
// BACKFILL_BATCH_BUFFER_SIZE = 5
if self.batches.iter().filter(...).count() >= BACKFILL_BATCH_BUFFER_SIZE {
return None; // ← REJECT
}
// ... later adds batch via Entry::Vacant(entry).insert(...)
```
Without the `>` being changed to a `>=` , we would allow 6. The same idea applies to all changes proposed.
Co-Authored-By: Antoine James <antoine@ethereum.org>
Co-Authored-By: Jimmy Chen <jimmy@sigmaprime.io>
Co-Authored-By: Jimmy Chen <jchen.tc@gmail.com>
#8547
Bump the version of `uuid` in our Cargo.toml to version `1` which removes `uuid 0.8` and unifies it across the workspace to version `1.19.0`.
Co-Authored-By: Mac L <mjladson@pm.me>
Fix the failure of the beacon-chain tests for phase0/altair, which now only runs nightly.
Just skip the payload invalidation tests, they don't make any sense prior to Bellatrix anyway.
Co-Authored-By: Michael Sproul <michael@sigmaprime.io>
Fix a bug in v8.1.0 whereby the VC times out continuously with:
> Feb 18 02:03:48.030 WARN Head service failed retrying starting next slot error: "Head monitoring stream error, node: 0, error: SseClient(Transport(reqwest::Error { kind: Decode, source: reqwest::Error { kind: Body, source: TimedOut } }))"
- Remove the existing timeout for the events API by using `Duration::MAX`. This is necessary as the client is configured with a default timeout. This is the only way to override/remove it.
- DO NOT add a `read_timeout` (yet), as this would need to be configured on a per-client basis. We do not want to create a new Client for every call as the early commits on this branch were doing, as this would bypass the TLS cert config, and is also wasteful.
Co-Authored-By: hopinheimer <knmanas6@gmail.com>
Co-Authored-By: Michael Sproul <michael@sigmaprime.io>
Co-Authored-By: Michael Sproul <michaelsproul@users.noreply.github.com>
Since https://github.com/sigp/lighthouse/pull/6847, invalid `BlocksByRange`/`BlobsByRange` requests, which do not comply with the spec, are [handled in the Handler](3d16d1080f/beacon_node/lighthouse_network/src/rpc/handler.rs (L880-L911)). Any peer that sends an invalid request is penalized and disconnected.
However, other kinds of invalid rpc request, which result in decoding errors, are just dropped. No penalty is applied and the connection with the peer remains.
I have added handling for the `ListenUpgradeError` event to notify the application of an `RPCError:InvalidData` error and disconnect to the peer that sent the invalid rpc request.
I also added tests for handling invalid rpc requests.
Co-Authored-By: ackintosh <sora.akatsuki@gmail.com>
None
I noticed that `observed_column_sidecars` is missing its prune call in the finalization handler, which results in a memory leak on long-running nodes (very slow (**7MB/day**)) :
13dfa9200f/beacon_node/beacon_chain/src/canonical_head.rs (L940-L959)
Both caches use the same generic type `ObservedDataSidecars<T>:`
22ec4b3271/beacon_node/beacon_chain/src/beacon_chain.rs (L413-L416)
The type's documentation explicitly requires manual pruning:
> "*The cache supports pruning based upon the finalized epoch. It does not automatically prune, you must call Self::prune manually.*"
b4704eab4a/beacon_node/beacon_chain/src/observed_data_sidecars.rs (L66-L74)
Currently:
- `observed_blob_sidecars` => pruned
- `observed_column_sidecars` => **NOT** pruned
Without pruning, the underlying HashMap accumulates entries indefinitely, causing continuous memory growth until the node restarts.
Co-Authored-By: Antoine James <antoine@ethereum.org>
