Repository navigation
feat(wal): add durable quorum append and recovery - #400
Merged
Merged
Conversation
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
A local WAL sync alone cannot justify an ingestion ACK. Add an optional, transport-independent raft-rs append driver over the shared durable log: persist Ready entries/HardState and LightReady commit updates before returning committed completions. Exact latest retries pass through consensus again and return the original accepted index. Leadership loss reports unresolved requests as unknown; index-plus-term matching prevents a replaced suffix from completing another request.
Input bodies remain shared
Bytesin Raft entries. Immutable layout/application-policy binding protects recovery. Accepted metadata, producer tracking and pending proposals have explicit limits; groups can move to independent workers sharing one engine's sync/failure fence. Accepted-record pages capture their cursor with their contents so a subsequent publisher cannot skip intervening appends by reading status separately.This increment retains the full log and static membership. Authenticated transport, bounded disk use/reclamation, producer lifecycle, snapshots/replica replacement and Lance publication integration remain separate work. It does not enable production ingestion or establish a throughput result. See
docs/design/replicated-wal-append.mdfor the API and durability boundaries.Refs #378
Testing