DanielLeens commented on PR #10551: URL: https://github.com/apache/seatunnel/pull/10551#issuecomment-6557490736
Thanks @SEZ9. The actual conflict and four failed OS/JDK unit lanes remain separate gates. The author's next source change should target the **existing reset/restore admission blocker** from my current-head review: `PhysicalVertex.reset` can return after a missing/refused/failed reset, while SubPlan continues installing fresh task futures and admits redeployment. That needs explicit reset success/failure propagation before a new run is published. I would not reopen the older cleanup-generation/deadlock list wholesale. The detailed review of this head already recognizes generation-bound retained cleanup records/retries and selective cleanup-before-state lock ordering; it did not substantiate a deadlock solely from JobMaster.init being under the job fence. The lock-order comment correction and isolated log-capture test remain the existing non-blocking maintenance items. Please preserve those implemented protections when resolving the conflict. The failed Build does not clear or replace the reset blocker. Its four unit matrices require a clean rerun on the eventual corrected/conflict-resolved head, with any repeated failure diagnosed from that attempt. No new full review or new numbered finding is being added on this unchanged head. Checked 2026-11-01T11:45:27.880206+00:00; unchanged head `61cd8a20a87e52cf42ec2701ba4b91128bd92e67`, base `dev`; compare `diverged`, ahead 18 / behind 65; merge `dirty` / `DIRTY`; review decision `REVIEW_REQUIRED`. Behind/diverged is recorded independently and does not itself request a sync. Daniel has READ repository permission. -- 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]
