databendlabs / openraft / 28119054731

87%

Build Type

push

github

Committed by drmingdrmer

Commit Message

fix: stop replication after storage read error

# Summary

Replication now treats a storage read error while generating
AppendEntries as fatal to the replication task, instead of letting
stream exhaustion look like a successful send and retrying the payload.

# Details

Add an integration regression test for issue #1795 that injects a
one-shot memstore read error while a leader replicates a live suffix
after a purged prefix. The test records whether the leader emits the
malformed retry with `prev_log_id=None` and non-empty entries.

Record a fatal marker in the stream state after reporting the storage
error to RaftCore. The owning replication loop checks this marker both
after stream setup and after response consumption so eager custom
streaming transports and the default lazy stream path both stop
immediately.

Add a debug assertion on the generated AppendEntries request to verify
that a non-empty entries vector starts exactly after its `prev_log_id`.

- Fixes #1795

Coverage Stats

41 of 45 new or added lines in 3 files covered. (91.11%)

6 existing lines in 3 files now uncovered.

16816 of 19354 relevant lines covered (86.89%)

166610.35 hits per line