88fantasy commented on PR #12388: URL: https://github.com/apache/seatunnel/pull/12388#issuecomment-6484930992
Thanks @DanielLeens for the reviews and CI follow-up. I have rechecked this Apache SeaTunnel PR against the latest dev. #12042 was merged on October 28, 2026 as b274eb1f2877f62a534933253c6243248521b7e2. It covers both parts of this fix: recognizing finished unbounded reader groups during running-pipeline restoration, and reconstructing closedIdleTask before the next checkpoint. It also includes stricter eligibility checks, unit coverage, and CheckpointCoordinatorFailoverIT.testCheckpointCompletesAfterMasterFailoverWithIdleClosedReader. I am closing #12388 as superseded by that merged implementation rather than retaining two competing restore paths. The associated issue #12386 is resolved by #12042. For clarity, this does not claim that this PR's CI is green: its JDK 17 engine-v2-it failure is a worker-loss/cancellation race (expected CANCELED, actual FAILED). The inspected log shows a reader transitioning from RUNNING to FAILED through CoordinatorService.makeTasksFailed, without a master takeover. That failure should be tracked separately; the related open #12311 is not a verified cure for this exact window. No fresh real-cluster acceptance of the merged implementation is claimed here. Thank you for the review and for landing the upstream fix. -- 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]
