DanielLeens commented on issue #12243: URL: https://github.com/apache/seatunnel/issues/12243#issuecomment-5635782859
Thank you for supplying the upstream reproduction and keeping #12264 test-only. This clears the earlier evidence gate for the fixed-bucket path. In the cited current source, `PaimonBucketAssigner` is used only for `HASH_DYNAMIC`; the fixed-bucket branch owns a table writer in each sink subtask, so this is not covered by the dynamic-routing check. #12264 is now the correct place to validate the reproducer, but its current Build check is failing. Please make that draft CI-clean before a production fix is discussed, and keep the first review focused on proving the two-writer failure and the single-writer control without changing production behavior. A follow-up fix must be a separate PR. It needs to state the ownership/commit contract for fixed buckets across parallel writers, preserve checkpoint and restore behavior, and cover the same two-writer cases plus recovery. Do not broaden `PaimonBucketAssigner` or replace the existing dynamic-routing path until the reproducer is green and the fixed-bucket ownership change is justified. -- 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]
