persistence.persistentVolumeClaim.trivy.existingClaim can't have value

[linonetwo]

Expected behavior and actual behavior:

Need to provide existingClaim when upgrade or change values using helm upgrade harbor --namespace image-registry ./harbor, otherwise there will be "device already mounted at" error:

MountVolume.SetUp failed for volume "pvc-4bcbe5a9-200f-44a3-bd73-101d7951589c" : rpc error: code = Internal desc = rpc error: code = Internal desc = verifyMount: device already mounted at [/var/lib/kubelet/pods/13dacfb6-b1e3-4037-9e92-2e67856858b4/volumes/kubernetes.io~csi/pvc-4bcbe5a9-200f-44a3-bd73-101d7951589c/mount /host/var/lib/kubelet/pods/13dacfb6-b1e3-4037-9e92-2e67856858b4/volumes/kubernetes.io~csi/pvc-4bcbe5a9-200f-44a3-bd73-101d7951589c/mount]

Steps to reproduce the problem:

But if fill value at persistence.persistentVolumeClaim.trivy.existingClaim, it also gives error:

helm upgrade harbor --namespace image-registry ./harbor
Error: UPGRADE FAILED: cannot patch "harbor-trivy" with kind StatefulSet: StatefulSet.apps "harbor-trivy" is invalid: spec: Forbidden: updates to statefulset spec for fields other than 'replicas', 'ordinals', 'template', 'updateStrategy', 'persistentVolumeClaimRetentionPolicy' and 'minReadySeconds' are forbidden

while fill this value for registry and jobservice is OK. it is not used and still cause "device already mounted at" for pod

persistence:
  enabled: true
  resourcePolicy: "keep"
  persistentVolumeClaim:
    registry:
      existingClaim: "harbor-registry"
      storageClass: ""
      subPath: ""
      accessMode: ReadWriteOnce
      size: 5Gi
      annotations: {}
    jobservice:
      jobLog:
        existingClaim: "harbor-jobservice"

Versions:

• harbor version: 2.10.2 / 1.14.2

Additional context:

[linonetwo]

Can be workaround by manually deleting the old pod when upgrading helm.

[MinerYang]

Will close by non-related.