DanielLeens commented on issue #12045: URL: https://github.com/apache/seatunnel/issues/12045#issuecomment-5509300621
Thanks for the report. I checked the exact 1d18735b tree and the #12013 merge. This test constructs its own CoordinatorService inside the test method; newMockCoordinatorService stops only the two auxiliary schedulers, and the coordinator executor is not shut down until this test reaches its own finally block. The production master-activation path already recreates that executor after an actual coordinator cleanup. Please do not make startPendingJobScheduleThread swallow a rejected submission or recreate executors on an arbitrary call, because that could hide a real lifecycle defect. The first PR should be test-only: make the executor lifecycle deterministic or observable, identify the shutdown caller, and remove that cause. If a real master step-down reproduces it, extend the regression through checkNewActiveMaster and clearCoordinatorService before proposing a production change. Please link the focused PR here. -- 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]
