Disk resizing isn't working. Changes are ignored.

[zebesh]

/kind bug

What happened?
We updated the driver to the latest version via helm upgrade (chart version 2.54.1)
After that, we started changing the size of the volume claim template and nothing happened. Before this, everything worked.
The status shows:
allocatedResourceStatuses:
storage: NodeResizePending

There are no resize messages in the controller logs.

What you expected to happen?
The driver should have increased the size of the volume claim template and restarted the pod.

How to reproduce it (as minimally and precisely as possible)?
Install driver 2.54.1 and change the size of the volume claim template.

Anything else we need to know?:

Environment

• Kubernetes version (use kubectl version): Server Version: v1.33.5-eks-3025e55
• Driver version: 1.54.0

[ConnorJC3]

Hi @zebesh, based on the information you provided it's likely you're running into this Kubernetes bug: kubernetes/kubernetes#131402

This was a bug in Kubernetes that occurs when the control plane is 1.32 or higher and the node version is 1.31 or lower.

If this bug is affecting your PVC, you need to take one of the following actions:

• Upgrade your nodes to Kubernetes 1.32 or higher
• Upgrade your nodes to a patch release with the fix included (See the above linked issue)
• (Strongly discouraged, use at your own risk) Disable the RecoverVolumeExpansionFailure on all EBS CSI Driver sidecars

Even after implementing the fix, PVCs affected by this bug may still be in a stuck state (the fix should prevent any new PVCs from getting stuck). In order to unstick them, you can manually clear the status.allocatedResourceStatuses field on the stuck PVCs.