Caideyipi commented on PR #18623:
URL: https://github.com/apache/iotdb/pull/18623#issuecomment-6620739833

   Re-review completed at `09bd6c2bf03c25d9824aad8e1a455948450e58ad` and found 
one behavioral issue in `LogicalBackupWriter`.
   
   When `forceIfNeeded()` failed after records had been appended, the writer 
left the segment tail and in-memory state in place. A sink retry could then 
create a duplicate event group, while a retried `SKIPPED_EVENT` could leave 
manifest counters inconsistent with the segment. The fix rolls back 
event/control writes when fsync or manifest persistence fails and updates the 
skipped-event counter inside the same rollback-protected transaction. It adds 
regression coverage for both retry cases.
   
   Verification:
   - Clean dependency-reactor regression run: 2/2 new tests passed.
   - Clean focused reactor: 51 archive/writer/sink tests passed (writer 24, 
archive reader 17, DataNode sink 8, ConfigNode sink 2).
   - Spotless and `git diff --check` passed; all PR-added tests have run 
locally.
   - POSIX-only permission assertions, the two-cluster integration suite, and 
full process-restart/leader-failover scenarios were not rerun in this round.
   
   Commit `5c9adc8a80c5ac176f8d780b840a28b31d3172b7` is pushed. Local HEAD, the 
source branch, and PR head all match this commit. The PR remains open and 
Draft; remote CI is not reported as passed.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to