Resizer logging with 'Failed to watch *v1beta1.VolumeAttributesClass when feature gate is enabled but API is unavailable'

[badrinathpathak]

What happened?

When running external-resizer with the VolumeAttributesClass feature gate set to true, the controller attempts to watch VolumeAttributesClass resources.

On clusters where this API is not available (e.g., OCP 4.19 / Kubernetes 1.32, where VAC is still beta and must be manually enabled), the resizer pod logs errors repeatedly:

Failed to watch *v1beta1.VolumeAttributesClass: failed to list *v1beta1.VolumeAttributesClass: the server could not find the requested resource" logger="UnhandledError"

This is confusing to operators and can be mistaken as a functional failure.

---

What you expected to happen?

• If the VolumeAttributesClass API is not available in the cluster, resizer should:

  • Detect this via API discovery,
  • Log a clear one-time warning,
  • Skip watching VAC resources,
  • Continue normal resizing functionality.

---

How to reproduce it (as minimally and precisely as possible)

1. Deploy external-resizer with:

   VolumeAttributesClass=true

2. Run on a cluster where VolumeAttributesClass is not enabled:

   oc api-resources | grep volumeattributesclasses
   # no output

3. Observe repeated error logs from the resizer pod.

---

Anything else we need to know?

• Workaround:

  • Ignore the error if VAC is not required, OR
  • Enable the VolumeAttributesClass API via feature gate.

---

Environment

• external-resizer version: v1.14.0
• Kubernetes version: v1.32
• Platform: OCP 4.19
• Feature gate: VolumeAttributesClass=true

[gnufied]

So, I think we are going to require k8s v1.34 at very minimum for using the newer resizer and hence this will be a non-issue for new release we are in process of making.

I agree that, this is an issue for older releases. So this will be strictly a fix that will go in older releases of external-resizer.