Add error rates metrics on unstable to benchmark against tree-sync. In my branch there are frequent errors but mostly connections errors as the node is still finding it set of stable peers.
These metrics are very useful and unstable can benefit from them ahead of tree-sync
Add three new metrics:
- sync_rpc_requests_success_total: Total count of sync RPC requests successes
- sync_rpc_requests_error_total: Total count of sync RPC requests errors
- sync_rpc_request_duration_sec: Time to complete a successful sync RPC requesst
Co-Authored-By: dapplion <35266934+dapplion@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>
N/A
Fixes the issue where we were setting block observed timings for blocks that were potentially gossip invalid.
Thanks @gitToki for the find
Co-Authored-By: Pawan Dhananjay <pawandhananjay@gmail.com>
Co-Authored-By: Michael Sproul <michaelsproul@users.noreply.github.com>
I accidentally broke `unstable` while merging some missed commits from `release-v8.0`. The merge was clean but semantically broken, and I didn't notice because I pushed without running CI 😬
- Fix the regression test added for #8528, for compatibility with the recent `RpcBlock` changes. I'm passing `is_available = false` which seems correct for this test.
Co-Authored-By: Michael Sproul <michael@sigmaprime.io>
Swaps out the `local_ip_address` dependency for `if-addrs`. The reason for this is that is that `local_ip_address` is a relatively heavy dependency (depends on `neli`) compared to `if-addrs` and we only use it to check the presence of an IPv6 interface. This is an experiment to see if we can use the more lightweight `if-addrs` instead.
Co-Authored-By: Mac L <mjladson@pm.me>
N/A
In https://github.com/sigp/lighthouse/pull/4801 , we added a state lru cache to avoid having too many states in memory which was a concern with 200mb+ states pre tree-states.
With https://github.com/sigp/lighthouse/pull/5891 , we made the overflow cache a simpler in memory lru cache that can only hold 32 pending states at the most and doesn't flush anything to disk. As noted in #5891, we can always fetch older blocks which never became available over rpc if they become available later.
Since we merged tree states, I don't think the state lru cache is relevant anymore. Instead of having the `DietAvailabilityPendingExecutedBlock` that stores only the state root, we can just store the full state in the `AvailabilityPendingExecutedBlock`.
Given entries in the cache can span max 1 epoch (cache size is 32), the underlying `BeaconState` objects in the cache share most of their memory. The state_lru_cache is one level of indirection that doesn't give us any benefit.
Please check me on this cc @dapplion
Co-Authored-By: Pawan Dhananjay <pawandhananjay@gmail.com>
Adds support for payload envelopes in the db. This is the minimum we'll need to store and fetch payloads.
Co-Authored-By: Eitan Seri- Levi <eserilev@gmail.com>
All the required boilerplate for gloas gossip. We'll include the gossip message processing logic in a separate PR
Co-Authored-By: Eitan Seri- Levi <eserilev@gmail.com>
Co-Authored-By: Jimmy Chen <jchen.tc@gmail.com>
#8652
Implements "simplified" versions of `max_blobs_by_root_request` and `max_data_columns_by_root_request` which do not depend on type information from the `data` module. I've also added tests which test the original implementation against the simplified one to ensure they don't deviate.
Also moves `all_data_column_sidecar_subnets` from a method on `ChainSpec` to a function which just takes `ChainSpec` as an argument.
Co-Authored-By: Mac L <mjladson@pm.me>
Pulling out consensus type changes from #8677.
This PR covers all type changes for spec 1.7.0-alpha.1 (except for `DataColumnSidecar` changes, which is covered in @eserilev's PR #8682)
Co-Authored-By: Jimmy Chen <jchen.tc@gmail.com>
Co-Authored-By: Michael Sproul <michael@sigmaprime.io>
Removes the remaining facade re-exports from `consensus/types`.
I have left `graffiti` as I think it has some utility so am leaning towards keeping it in the final API design.
Co-Authored-By: Mac L <mjladson@pm.me>
Removes some of the temporary re-exports in `consensus/types`.
I am doing this in multiple parts to keep each diff small.
Co-Authored-By: Mac L <mjladson@pm.me>
#8652
- This removes instances of `BeaconStateError` from `eth_spec.rs`, and replaces them directly with `ArithError` which can be trivially converted back to `BeaconStateError` at the call site.
- Also moves the state related methods on `ChainSpec` to be methods on `BeaconState` instead. I think this might be a more natural place for them to exist anyway.
Co-Authored-By: Mac L <mjladson@pm.me>
N/A
The `beacon_data_column_sidecar_computation_seconds` used to record the full kzg proof generation times before we changed getBlobsV2 to just return the full proofs + cells. This metric should be taking way less time than 100ms which was the minimum bucket previously.
Update the metric to use the default buckets for better granularity.
Co-Authored-By: Pawan Dhananjay <pawandhananjay@gmail.com>
Just visual clean-up, making logging statements look uniform. There's no reason to use `tracing::debug` instead of `debug`. If we ever need to migrate our logging lib in the future it would make things easier too.
Co-Authored-By: dapplion <35266934+dapplion@users.noreply.github.com>
Co-Authored-By: Jimmy Chen <jchen.tc@gmail.com>
Co-Authored-By: Michael Sproul <michaelsproul@users.noreply.github.com>
Which issue # does this PR address?
None
Discussed in private with @jimmygchen, Lighthouse's `earliest_available_slot` is guaranteed to always align with epoch boundaries, but as a safety implementation, we should use `start_slot` just in case other clients differ in their implementations.
At least we agreed it would be safer for `synced_peers_for_epoch`, I also made the change in `has_good_custody_range_sync_peer`, but this is to be reviewed please.
Co-Authored-By: Antoine James <antoine@ethereum.org>
Co-Authored-By: Jimmy Chen <jimmy@sigmaprime.io>
```bash
$ lcli mock-el ....
...
...
Dec 15 11:52:06.002 INFO Metrics HTTP server started listen_address: "127.0.0.1:8551"
...
```
The log message "Metrics HTTP server" was misleading, as the server is actually a Mock Execution Client that provides a JSON-RPC API for testing purposes, not a metrics server.
Co-Authored-By: ackintosh <sora.akatsuki@gmail.com>
Closes#8569
Updates the HTTP API error when the node cannot reconstruct blobs due to "Insufficient data columns".
Changes the response from 500 Internal Server Error to 400 Bad Request and adds a hint to run with --supernode or --semi-supernode.
Co-Authored-By: Andrurachi <andruvrch@gmail.com>
Which issue # does this PR address?
#8586
Please list or describe the changes introduced by this PR.
Remove `service_name` from `TaskExecutor`
Co-Authored-By: Abhivansh <31abhivanshj@gmail.com>
While reviewing Gloas I noticed we were updating `PartialBeaconState`. This code isn't used since v7.1.0 introduced hdiffs, so we can delete it and stop maintaining it 🎉
Similarly the `chunked_vector`/`chunked_iter` code can also go!
Co-Authored-By: Michael Sproul <michael@sigmaprime.io>
Co-Authored-By: Pawan Dhananjay <pawandhananjay@gmail.com>
Which issue # does this PR address?
None
The `visualize_batch_state` functions uses the following loop `for mut batch_index in 0..BATCH_BUFFER_SIZE`, making it from `0` to `BATCH_BUFFER_SIZE - 1` (behind the scenes).
Hence we would never hit the following condition:
```rust
if batch_index != BATCH_BUFFER_SIZE {
visualization_string.push(',');
}
```
Replacing `!=` with `<` & `BATCH_BUFFER_SIZE -1` allows for the following change:
`[A,B,C,D,E,]` to become: `[A,B,C,D,E]`
Co-Authored-By: Antoine James <antoine@ethereum.org>
#8547
Update our `strum` dependency to `0.27`. This unifies our strum dependencies and removes our duplication of `strum` (and by extension, `strum_macros`).
Co-Authored-By: Mac L <mjladson@pm.me>
Co-Authored-By: Michael Sproul <michaelsproul@users.noreply.github.com>
Closes#8426
Added a new regression test: `reproduction_unaligned_checkpoint_sync_pruned_payload`.
This test reproduces the bug where unaligned checkpoint syncs (skipped slots at epoch boundaries) fail to import the anchor block's execution payload when `prune_payloads` is enabled.
The test simulates the failure mode by:
- Skipping if execution payloads are not applicable.
- Creating a harness with an unaligned checkpoint (gap of 3 slots).
- Configuring the client with prune_payloads = true.
It asserts that the Beacon Chain builds successfully (previously it panicked with `MissingFullBlockExecutionPayloadPruned`), confirming the fix logic in `try_get_full_block`.
Co-Authored-By: Andrurachi <andruvrch@gmail.com>
Co-Authored-By: Michael Sproul <michaelsproul@users.noreply.github.com>
