SEZ9 commented on PR #12943: URL: https://github.com/apache/seatunnel/pull/12943#issuecomment-6842819142
Thanks for digging into the logs on `16714019` — a mixed DNS/IP member identity after the active-master replacement is a much more satisfying explanation than "CI is slow", and it would explain why both JDK 17 attempts (`129705577600` and `130364594639`) hit the same post-deletion wait at different assertions while `4504ff81` / `713de07d` were green. A few things so I can close this out: 1. **Please push the change (or point me to the new head).** The last head I see here is `16714019`, which only merged `dev`, so I can't see the Pod-IP advertisement or the staged recovery checks yet. 2. **Scope check.** Advertising the current Pod IP while keeping stable DNS names for discovery sounds like a change to the HA master's network/member configuration, not only to the test. Could you briefly describe where that lands (deployment manifests vs. engine config vs. the IT itself) so I can confirm the deployment-only HA scope from the earlier review still holds, or re-review the new surface if it doesn't? 3. **Staged checks.** Glad you kept the 180 s budget and made the failing stage identifiable — that's essentially Option B. Please confirm the failure message now names which of the three conditions (replacement Pod `Running` → job back to `RUNNING` → promoted master visible to the probe) was still false, so a future failure is immediately actionable. 4. **Evidence.** Once pushed, I'd like to see `engine-resource-managers-it (17)` pass on the new head, ideally with at least one re-run so we have more than a single green data point for the `loseActive=true` cases and `applicationIdCancelRemovesRunningRecoveredHaApplication`. Nothing else is outstanding from my side beyond the above; I'll re-check as soon as the new head is up. <!-- streview-comment:2313 --> -- 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]
