"kapp: Error: update customresourcedefinition/volumegroupsnapshotclasses" Reconcile failed for a VKS Cluster's CSI package post VKR upgrade from 1.31 to 1.32 or higher
book
Article ID: 407835
calendar_today
Updated On:
Products
VMware vSphere Kubernetes Service
Issue/Introduction
Post VKS Cluster upgrade from 1.31 to 1.32, one package(pkgi) "<workload cluster name>-vsphere-pv-csi" is in "Reconcile failed" state
While describing the failed pkgi, similar log entries will be seen
k describe pkgi -n vmware-system-tkg "<workload cluster name>-vsphere-pv-csi"
Useful Error Message: kapp: Error: update customresourcedefinition/volumegroupsnapshotclasses.groupsnapshot.storage.k8s.io(apiextensions.k8s.io/v1) cluster: Updating resource customresourcedefinition/volumegroupsnapshotclasses.groupsnapshot.storage.k8s.io (apiextensions.k8s.io/v1) cluster: API server says: CustomResourceDefinition.apiextensions.k8s.io "volumegroupsnapshotclasses.groupsnapshot.storage.k8s.io" is invalid: status.storedVersions[0]: Invalid value: "v1alpha1": must appear in spec.versions (reason: Invalid)
Environment
vSphere Supervisor 9.x
vSphere Supervisor 8.x
VKS Service 3.2.0 and higher
Workload Cluster upgrading to KR v1.31 or higher
Cause
CRD volumegroupsnapshotclassesshipped with pv-csi v3.4.0 have version v1beta1.
CRD volumegroupsnapshotclasses.groupsnapshot.storage.k8s.io reported have version v1alpha1
Resolution
Validate if the CRD is in use by running below command from Guest Cluster Context
kubectl get volumegroupsnapshotclasses --all-namespace
If the command returns no output, it indicates that no resources are using this CRD. [Do not proceed to the next step if there are any resources in use, contact Broadcom Support Team]
The CRD can be safely deleted using below commands