[FLINK-16945] Execute CheckpointFailureManager.FailJobCallback directly in main thread executor

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.10.0, 1.11.2, 1.12.0
• Fix Version/s: 1.12.0
• Component/s: Runtime / Checkpointing
• Labels: pull-request-available stale-assigned
• Assignee: Biao Liu
• Reporter: Biao Liu
• Created: 02/Apr/20 12:10
• Updated: 22/Jun/21 13:55
• Resolved: 23/Apr/21 06:50

Description

Since we have put all non-IO operations of CheckpointCoordinator into main thread executor, the CheckpointFailureManager.FailJobCallback could be executed directly now. In this way execution graph would fail immediately when CheckpointFailureManager invokes the callback. We could avoid the inconsistent scenario of FLINK-13497.

Issue Links

• causes: FLINK-13497 Checkpoints can complete after CheckpointFailureManager fails job
• fixes: FLINK-13497 Checkpoints can complete after CheckpointFailureManager fails job
• links to: GitHub Pull Request #11648