Locust-based becnhmarking framework#34
Draft
dmitriy-b wants to merge 23 commits into
Draft
Conversation
…nap-sync + replay Post-merge EL clients only snap-sync or execute blocks when a CL drives their Engine API. Frozen milestone snapshots (the bloated states in #21) have no live CL, so lighthouse can't drive them. This adds a minimal Engine-API mock CL with the two CL roles #21 needs: - PivotDriver: repeated engine_forkchoiceUpdatedV3 to a fixed frozen pivot hash -> the EL snap-syncs to that milestone state (snap-sync metrics). - ReplayDriver: engine_newPayloadV{1..4} + engine_forkchoiceUpdated over recorded payloads -> drives execution under live head + per-block latency (block-processing p50/p95/p99, gas/s metrics). EngineClient hand-rolls HS256 JWT (no new deps) and auto-picks newPayload/FCU versions across Paris..Prague. Wired into the orchestrator via consensus_mode: lighthouse|mock|none (back-compat: legacy consensus bool still works). Generalizes the throwaway geth FCU/replay scripts. Nethermind targets can use --Sync.StaticSnapPivot (nethermind#11943) as the native pivot alternative.
Mock consensus layer (Engine-API driver) for frozen-milestone snap-sync + replay
- scripts/rlp_to_mocks.py: the RLP-stream->JSONL converter was orphaned from commit 8177dcf (its test was committed without the module). Add it so the suite resolves. - templates/neth/: self-contained Nethermind x3.5 bloatnet stack mirroring templates/geth/ — chainId-1 shadowfork with --Init.NetworkId=12159, FlatDb, mounted chainspec, ports clear of a co-located host-net geth (RPC 8547 / engine 8552), and CADVISOR_IMAGE pinned to the containerd-snapshotter fix. Verified end-to-end on state-vm: replay drives the head and the full runner (DB-size + Locust + aggregator) runs against it. - templates/geth/: track the existing self-contained geth template too. - README: document scripts/rlp_to_mocks.py, the per-client templates/ stacks, the cAdvisor containerd-snapshotter fix, and the replay-driven load-test state-window caveat. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The x3.5 FlatDb node served historical state only ~128 blocks deep (the FlatDb.MinReorgDepth default), so a load test that pins recent blocks while the mock-CL replay advances the head hit -32002 "No state available" once pinned blocks aged out — Locust then exited non-zero and the runner skipped the aggregator. Raise --FlatDb.MinReorgDepth=1024 / --FlatDb.MaxReorgDepth=2048 (above load_test.recent_block_window=1000). Verified on state-vm: full run (DB-size + Locust 5m + aggregator) with concurrent replay → 47,127 requests, 0 failures (was ~10.6%), metrics_nethermind.json produced with full cAdvisor resource metrics. README updated with the FlatDb depth guidance. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
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.
No description provided.