[FLINK-25974] Make cancellation of jobs depend on the JobResultStore

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.15.0
• Fix Version/s: 1.15.0
• Component/s: Runtime / Coordination
• Labels: pull-request-available
• Assignee: Matthias Pohl
• Reporter: Matthias Pohl
• Created: 07/Feb/22 10:46
• Updated: 14/Feb/22 08:08
• Resolved: 14/Feb/22 07:47

Description

JobManagerRunner instances were cancellable as long as the instance was still registered in the Dispatcher.jobManagerRunnerRegistry. With the cleanup being done concurrently (i.e. not relying on the JobManagerRunnerRegistry to be cleaned up anymore), the cancellation of a job should only be possible as long as the job is not globally finished and before cleanup is triggered.

We should verify whether a job is listed in the JobResultStore and only enable the user to cancel the job if that's not the case.

Issue Links

• Testing discovered: FLINK-26114 DefaultScheduler fails fatally in case of an error when shutting down the checkpoint-related resources
• links to: GitHub Pull Request #18644