PVC expansion failed on AWS

[jmcorne]

Please check the FAQ documentation before raising an issue

Describe the bug (required)

When you increase the size of PVCs (metad or storaged ), the operator failed to process with the following error.

    nebula_cluster_control.go:152] reconcile storaged cluster failed: 
      storageclasses.storage.k8s.io is forbidden: User "system:serviceaccount:<my_namespace>:nebula-controller-manager-sa" 
      cannot list resource "storageclasses" in API group "storage.k8s.io" at the cluster scope

Your Environments (required)

• OS: uname -a: Linux nebula-controller-manager-deployment-6d94bc886-66st4 5.10.238-234.956.amzn2.x86_64 first commit #1 SMP Tue Jul 1 20:20:57 UTC 2025 x86_64 Linux
• Commit id (e.g. a3ffc7d8): ??
• Nebula Operator tag: vesoft/nebula-operator:v1.8.3
• Nebula Operator SHA: docker.io/vesoft/nebula-operator@sha256:c1d39f1c4253d01dad97ac21b406bf9cf13d351af83ad6e506b5dffcdc939d0a

How To Reproduce(required)

Steps to reproduce the behavior:

1. Execute helm install with default values for the nebula-cluster Helm chart
2. Wait for completion
3. Then override a value (nebula.metad.dataVolume.storage) from "2Gi" to "4Gi"
4. Do a helm upgrade --set nebula.metad.dataVolume.storage="4Gi"
5. Check the Nebula Operator logs and size of PVC remains to 2Gi

Note that the storage class used in

AWS storage class is gp3 and is marked as expandable.

Expected behavior
Nebula Operator increases the size of the PVC to the desired sized.

Additional context

[jmcorne]

I did a workaround by patching the controller manager cluster role

- apiGroups:
    - storage.k8s.io
  resources:
    - storageclasses
  verbs:
    - list
    - get

And PVC expansion works as expected