SEZ9 commented on issue #11679: URL: https://github.com/apache/seatunnel/issues/11679#issuecomment-5225135723
Thanks for the detailed reproduction and for confirming this exists on pristine `dev` independent of #11648. The key symptoms to investigate: (a) SubPlan.handleCheckpointError repeatedly resubmitting a pipeline that fails deterministically (no retry cap / non-retryable error classification for authorization failures), and (b) the job vanishing from coordinator counters without ever transitioning to FAILED/CANCELED, so the client-side job-completion future never completes. Could you attach the full coordinator and client logs around the cancellation (especially JobMaster/SubPlan state transitions and the client's waitForJobComplete path)? That would help confirm whether the job's end state is lost during pipeline restore or whether the client's listener/future is dropped after resubmission. A likely fix direction: treat NoPrivilegeException (and similar deterministic failures) as non-recoverable so the pipeline transitions to FAILED instead of being resubmitted, and ensure JobMaster always publishes a terminal JobResult to the client even when the pipeline is cancelled during checkpoint error handling. A client-side timeout/watchdog could be a defensive secondary fix but shouldn't be the primary one. <!-- streview-comment:72 --> -- 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]
