[jira] [Created] (FLINK-17869) Fix the race condition of aborting unaligned checkpoint

[Shang Yuanchun (Jira)]

Zhijiang created FLINK-17869:
--------------------------------

             Summary: Fix the race condition of aborting unaligned checkpoint
                 Key: FLINK-17869
                 URL: https://issues.apache.org/jira/browse/FLINK-17869
             Project: Flink
          Issue Type: Bug
          Components: Runtime / Checkpointing
            Reporter: Zhijiang
            Assignee: Zhijiang
             Fix For: 1.11.0


On ChannelStateWriter side, the lifecycle of checkpoint should be as follows:

start -> in progress/abort -> stop.

We must guarantee that #abort should be queued after #start, otherwise the aborted checkpoint might be started later again in the case of race condition.

There are two cases might trigger abort checkpoint:
 * One is CheckpointBarrierUnaligner#processEndOfPartition, which should abort all the current and future checkpoints, no need to judge the condition `isCheckpointPending()` as current code did. 
 * Another is CheckpointBarrierUnaligner#processCancellationBarrier, which should only abort the respective checkpoint id if already triggered before.

The unaligned checkpoint might be triggered either by task thread or netty thread inside ThreadSafeUnaligner. Anyway we should know the current triggered checkpoint id in order to handle both above cases properly.

Another bug is that during ChannelStateWriterImpl#abort, we should not remove the respective ChannelStateWriteResult. Otherwise it would throw IllegalArgumentException when ChannelStateWriterImpl#getWriteResult in the process of checkpoint. ChannelStateWriteResult should be created at #start method and only removed at #stop method.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)