DanielLeens commented on PR #13111: URL: https://github.com/apache/seatunnel/pull/13111#issuecomment-6628321082
Thanks @wgzhao, a single follow-up PR is fine with me. Please keep head `5b08306819` as it is: both items are non-blocking, my approval stands on this exact head, and rerunning the roughly 75-minute Doris IT for them is not worth it here. I agree with how you split it. It also makes sense to put the dual abort and any hash change in the same follow-up, since each of them changes the label prefix. For the follow-up: - **Upgrade window:** when `restoredLabelPrefix != null && !restoredLabelPrefix.equals(labelPrefix)`, add the extra `abortPreCommit(labelPrefix, lastCheckpointId + 1)` after the restored-prefix probe in `DorisSinkWriter.java:535-538`. Then extend `testRestoreAbortsTwoPhaseCommitLabelsWithThePersistedPrefix` so it asserts both prefixes are probed. With the extra call missing, that test should fail. - **CRC32:** if you switch hashes, the follow-up is the right place, so that operators only see one more prefix transition. If you keep CRC32, one FAQ sentence about very large many-to-one table sets (`docs/en` and `docs/zh`) is enough. On this PR, CI is green on `5b08306819` (apache-side `Build` passes, and fork run 41749530741 succeeded). It now only needs a committer's review and merge. Thanks again for the careful work on #13108. -- 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]
