afedulov opened a new pull request, #24195:
URL: https://github.com/apache/flink/pull/24195

   `CompactCoordinator` accumulates data in `currentInputFiles` and only rolls 
them into `inputFiles` when `snapshotState()` gets called. At the same time it 
relies on separately receiving checkpoint indications from the upstream 
operator via `processElement()` (`EndCheckpoint`). If the job terminates, the 
final `EndCheckpoint` can arrive before the `snapshotState()` gets called. This 
leads to data loss (all events in `currentInputFiles` get discarded).
   
   This is a backport from the 1.19 branch: 
https://github.com/apache/flink/pull/24120


-- 
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