SEZ9 commented on PR #12755: URL: https://github.com/apache/seatunnel/pull/12755#issuecomment-6357116296
Thanks for syncing with the latest `dev`. The new head `4c3f68db66` is a merge of the `dev` tip `4dcfa28ccc`, and the PR's own change set is the same as the one reviewed at `c6246e0ced`, so there is no new diff to re-review. Based on the previous review, the lite-member retry path already unblacklists only temporary entries, resets a stale master address before `continue`, and refreshes an empty discovery list, and the en/zh deployment notes are part of the change set. A few points I'd still like to close out: - **All-permanent blacklist**: when every candidate is permanently blacklisted, does the lite-member branch still wait out `hazelcast.max.join.seconds` with nothing to retry, and does the log message reflect that? If so, falling through to the original behaviour or adjusting the log would help. - **Blacklist log churn**: the joiner's own lite-member logs now go to FINE. Could you confirm whether Hazelcast's add/remove-blacklist INFO messages are still emitted on every retry cycle? - **Worker leak on the timeout path in the tests**: please confirm the started-before-master worker is shut down when the join times out, so a failure doesn't affect later tests in the same JVM. - **PR description / test Javadoc vs. tests**: please remove (or add) the listed test that doesn't exist and align the scenario numbering between the description and the Javadoc. Once these are addressed I'm happy to approve. <!-- streview-comment:1971 --> -- 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]
