[FLINK-18063] Fix the race condition for aborting checkpoint in CheckpointBarrierUnaligner

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 1.11.0
• Fix Version/s: 1.11.0, 1.12.0
• Component/s: Runtime / Checkpointing
• Labels: pull-request-available
• Assignee: Zhijiang
• Reporter: Zhijiang
• Created: 02/Jun/20 09:07
• Updated: 09/Jun/20 16:45
• Resolved: 09/Jun/20 09:49

Description

There are three aborting scenarios which might encounter race condition:

    1. CheckpointBarrierUnaligner#processCancellationBarrier

    2. CheckpointBarrierUnaligner#processEndOfPartition

    3. AlternatingCheckpointBarrierHandler#processBarrier

We only consider the pending checkpoint triggered by #processBarrier from task thread to abort it. Actually the checkpoint might also be triggered by #notifyBarrierReceived from netty thread in race condition, so we should also handle properly to abort it.

Issue Links

• links to: GitHub Pull Request #12460
• links to: GitHub Pull Request #12511