nzw921rx opened a new pull request, #10865: URL: https://github.com/apache/seatunnel/pull/10865
### Purpose of this pull request In a split-cluster (master failover) scenario, the pending job schedule thread was never restarted after a standby node became the new master. This caused pending jobs to remain stuck indefinitely after failover. **Changes**: - `CoordinatorService`: extract executor instantiation into `createCoordinatorExecutor()` to reuse consistent config on both init and failover paths - `CoordinatorService`: move `startPendingJobScheduleThread()` call into `masterActiveListener` so it fires on every master promotion - `CoordinatorService`: fix schedule loop exit condition from `while (true)` to `while (isActive)` so the thread stops cleanly on demotion - `CoordinatorService`: set `isActive = false` at the start of `shutdown()` to stop the schedule loop before tearing down resources - `CoordinatorService`: check both `isShutdown()` and `isTerminated()` before recreating the executor on failover - `CoordinatorServiceTest`: add unit tests covering the fixed lifecycle paths - `SplitClusterPendingJobLifecycleFailoverIT` + config: add E2E test verifying pending job scheduling survives master failover in a split cluster ### Does this PR introduce _any_ user-facing change? No. ### How was this patch tested? - `CoordinatorServiceTest`: unit tests covering executor recreation, schedule thread restart, and `isActive` flag transitions - `SplitClusterPendingJobLifecycleFailoverIT`: E2E test that triggers master failover and asserts pending streaming jobs are picked up by the new master -- 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]
