Stateless Flink State Snapshot

[Yanisdje]

Problem

When a FlinkStateSnapshot CR is deleted before the FlinkStateSnapshotController has reconciled it, the cleanup() method throws a NullPointerException on getStatus().getState(). The exception is caught in a way that prevents the finalizer from being removed, causing the CR to be permanently stuck in a terminating state. This blocks namespace deletion.

Root Cause

FlinkStateSnapshotController.cleanup() and FlinkResourceContextFactory.getFlinkStateSnapshotContext() assume the status is non-null, but FlinkStateSnapshot CRs are created without a status (the status subresource is only populated when reconcile() runs). If the CR receives a deletion timestamp before reconcile() runs, JOSDK calls cleanup() directly — the only code path that initializes null status is in reconcile(), not cleanup().

The NPE propagates to JOSDK (or is caught by the controller's catch block returning noFinalizerRemoval()), causing an infinite retry loop where every attempt crashes the same way.

reconcile() already handles this case correctly:

// status might be null here
flinkStateSnapshot.setStatus(
        Objects.requireNonNullElseGet(
                flinkStateSnapshot.getStatus(), FlinkStateSnapshotStatus::new));

cleanup() and updateErrorStatus() are missing this guard.

Impact

• FlinkStateSnapshot CRs with null status accumulate as zombie resources with finalizers that can never be cleared.
• Namespace deletion is blocked indefinitely

Fix

Add the same null-status initialization to cleanup() and updateErrorStatus() that already exists in reconcile(). A null-status snapshot was never triggered against Flink, so no data exists on storage and cleanup can safely proceed

[gyfora]

Is this change related?

[vsantwana]

Hello @Yanisdje
Can I please request you to add the JIRA ticket number on the PR so that it shows up on the FLINK JIRA here and also possibly get the ticket assigned to yourself before creating a PR (This is also highlighted in the contribution guide)
Asking because it allows other contributors to know the accurate state of the ticket and avoids duplicate work/interest.
Thanks!

[gyfora]

@Yanisdje are you still working on this? :)

[nihar4276]

@Yanisdje @gyfora We hit this bug at Bloomberg with flink kubernetes operator 1.14.0. If Yanisdje is unavailable, I can make a quick PR. Let me know, thanks

[gyfora]

Thanks @nihar4276 that would be great. I think you can for the existing work on this PR to help wrap it up and remove anything unnecessary :)

[Yanisdje]

Hello @gyfora and @nihar4276!

Apologies for the delayed response, it seems that I was receiving the notifications in the wrong email. I could get back to it tomorrow, will keep folks updated.

Thanks!

[nihar4276]

Nice to meet you both 💯 . I will let Yanis take this PR to the finish line, and feel free to ping me if you need help. Thanks :)