SEZ9 commented on PR #10808: URL: https://github.com/apache/seatunnel/pull/10808#issuecomment-5806603384
Thanks for the detailed source-level walkthrough on `0d4d8624c`. **F1/F2:** The explanation makes sense: no fallback when Hazelcast has not reported a master, the "first non-lite member" inference only when the confirmed master is a lite member, and the `(best effort)` wording printed to the operator. Before I close these, could you point me at the actual diff hunks for `getActiveMasterAddress()` / `describeActiveMasterResolution()` and the two tests you mention (`testUnknownMasterDoesNotSelectFallbackCoordinator`, `testInferredCoordinatorIsMarkedBestEffort`) so I can verify against the code rather than the description? One small ask on top: the "can lag behind the cluster during failover" caveat should also appear in the user-facing docs (see F7), not only in the Javadoc. **F6:** The explicit `Active master: UNKNOWN` note under the table plus `testUnknownCoordinatorIsReportedExplicitly` sounds like the right fix. Same request: a link to the diff for that note and test and I'll mark it resolved. **F3 / F4 / F5 / F7 / F8:** You mentioned pointers for F3/F4/F5/F8 in an earlier comment; I may have missed them. Could you restate them here with the specific doc section or diff hunk for each? - F3 / F7: the incompatible-changes entry that covers the changed `ACTIVE MASTER` semantics of the CLI member list in separated clusters and the standby/failover window where no node reports as active coordinator. - F4: the value of the "bounded timeout" for the final cancel-time metrics flush, and whether it is configurable. - F5: a short note in the REST API v1 doc that `nodeRole`/`coordinator`/`worker` expose cluster topology on an endpoint that is unauthenticated by default. - F8: the inline `java.util.Collections` FQN replaced with an import. Once I have those references I'll re-check and close out the remaining items. <!-- streview-comment:1277 --> -- 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]
