The GitHub Actions job "Flink Kubernetes Operator CI" on flink-kubernetes-operator.git/nihar4276.flink-39356 has failed. Run started by GitHub user nihar4276 (triggered by nihar4276).
Head commit for run: fb86c902c7197ae9a4731264b3c67ae92ee04343 / nihar4276 <[email protected]> [FLINK-39356] Fix NPE during FlinkStateSnapshot cleanup when status is null Problem: A FlinkStateSnapshot deleted before the controller ever reconciled it has no status, since the status is only initialized in reconcile(). The cleanup path builds a FlinkStateSnapshotContext, which dereferences the status via getStatus().toBuilder(), so cleanup() threw a NullPointerException before entering its try block. The finalizer was never removed, leaving the resource stuck in terminating state and blocking namespace deletion, with the failed cleanup retried indefinitely. Fix: Return DeleteControl.defaultDelete() from cleanup() when the status is null. Such a snapshot was never triggered against Flink, so there is nothing to dispose of and the finalizer can be removed immediately. Verification: Added FlinkStateSnapshotControllerTest#testCleanupWithNullStatus, covering both a savepoint and a checkpoint whose status is null. Verified manually on a local minikube cluster: a FlinkStateSnapshot deleted before the operator reconciles it now has its finalizer removed and the resource disappears, instead of hanging in terminating state. Report URL: https://github.com/apache/flink-kubernetes-operator/actions/runs/30974216968 With regards, GitHub Actions via GitBox
