Component removal for VCF Automation failed while executing KB441333 procedures

Products

• VCF Automation

Issue/Introduction

- Failed to remove VCFA component refer to KB441333 and below error report in lifecycle task:

"Unable to unregister components from VCF Operations"

- Workload pods entered ImagePullBackOff because local registry endpoint 198.18.128.16:5000 refused connections.

"curl: (7) Failed to connect to 198.18.128.16 port 5000 after 0 ms: Could not connect to server"

- Inspection of namespace vmsp-platform showed zot-1 and zot-2 in 0/1 READY state

- Describing the zot pods revealed volume attachment failures within the vSphere CSI driver interface.

"attachdetach-controller AttachVolume.Attach failed for volume "pvc-xxxxxxxxxxxxxx": timed out waiting for external-attacher of csi.vsphere.vmware.com CSI driver to attach volume"

 

Environment

VCF Automation 9.1

Cause

Unconsolidated virtual machine snapshots on the VCFA appliance locked the virtual disk chain, preventing the vSphere CSI driver from attaching persistent volumes to the Zot image registry pods (zot-1, zot-2)

Resolution

1.Delete all active snapshots on the VCFA virtual machine via vCenter Server to consolidate the VMDK disk chain.

2.Confirm volume attachment completes and verify zot-1 and zot-2 pods transition to 1/1 Running state in namespace vmsp-platform.

kubectl get pods -n vmsp-platform -l app.kubernetes.io/name=zot -w

3.Clear all stale pending and queued tasks in the vmsp-platform namespace.

kubectl delete task -n vmsp-platform --all

4.Verify that all VCFA application pods return to Running status

kubectl get pods -A | grep -vE 'Running|Completed'

5.Re-execute the VCFA component removal workflow per KB441333.