SEZ9 commented on issue #12509: URL: https://github.com/apache/seatunnel/issues/12509#issuecomment-5862383682
Thanks for the detailed report and reproduction. Could you share the Master/Worker logs around the failed restore (including the WrongTargetSlotException stack trace and the preceding resource-allocation/release messages), plus the job config and cluster configuration (slot count, checkpoint interval) so maintainers can confirm the sequence? Since you already have a focused unit regression on dev `146a1b5c5` and a minimal fix (abort deployment when `preApplyResources` returns false and let the bounded retry path recover), would you be willing to open a PR with both the test and the fix? Please also describe how the pre-applied resource futures are handled on the failure path so the retry does not reuse stale slot information. To confirm the engine-level root cause independently of DuckLake/S3, could you verify the same behavior with a simpler sink (e.g. Console or a local JDBC sink) under the same kill-worker/replace-worker scenario? <!-- streview-comment:1370 --> -- 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]
