SEZ9 commented on PR #11602: URL: https://github.com/apache/seatunnel/pull/11602#issuecomment-5674221751
Thanks @tomatotomata for tracing the two failures back to the previous head rather than just re-running them. The change looks right to me: treating a member that has already left the cluster as unavailable and omitting it from the terminal snapshot, instead of spending the retry budget on a node that can never respond, keeps the strict terminal behaviour for the genuinely transient cases while narrowing the skip-and-log path to the departed-member branch. Supplying the metrics timeout config and the Hazelcast flake-id dependency to the `JobMasterTest` fixtures, plus the departed-worker regression, is the correct way to close those fixture failures rather than papering over them. One non-blocking question: you mentioned the focused local run was blocked in `seatunnel-config-shade` by compile errors before `JobMasterTest` executed. Do those reproduce on a clean `dev` checkout with your JDK, or only on this branch? If the latter, please paste the first error so we can tell whether it is a real regression or a local environment issue. Otherwise I have nothing further on `a283df0d52`. <!-- streview-comment:1063 --> -- 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]
