tomatotomata commented on PR #11602: URL: https://github.com/apache/seatunnel/pull/11602#issuecomment-5185639861
I traced the terminal cleanup path you outlined and agree that the final-history call should not silently reuse the best-effort realtime result. I?m thinking of separating final collection from `getCurrJobMetrics`, so a timeout or partial worker response propagates back through `subPlanDone` and leaves the task-group context available for the existing retry, while realtime polling keeps its bounded best-effort behavior. I also want to add a focused regression test around timeout, persistence, and cleanup before changing the lifecycle code. The current branch is still two commits behind `dev`, so I?m folding the design check and that sync into the next update. Let me know if you think a dedicated final-collection method fits better than returning an explicit completeness result. -- 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]
