The VMSP runtime cluster failed to update to version 9.1.0.1 due to pre-health check issue with is related to stale metadata in the template

Products

• VMware Cloud Foundation

Issue/Introduction

- The health check failed during the Fleet lifecycle update pre-check with the following error: unable to validate vSphere objects: unable to resolve datastore identifier

- The VMSP package description referenced a deleted datastore and stale vm Template ID.

- After update the correct reference information in package description, the Replacement VMs deployed during reconciliation failed to fetch IP addresses because the template was hardcoded to a legacy hostname.

- Similar issue could be also found in VCFA upgrade precheck

Environment

VCF Management Services Runtime 9.1.x

VCF Automation 9.1.x

Resolution

1.Manual configuration of hostnames for new deployed node in the vCenter inventory refer to the output of kubectl get node to allowed the FluxCD reconciliation to move past the blocking stage.

2.After the reconciliation completed, remove the cloned template and update new one refer to KB 434922

3.Update PackageDeployment metadata (datastore, datastoreId, datastoreURL, and Template Id).

4.Delete the legacy storage class and apply a new storage class using the corrected datastoreURL.

5.Watch reconciliation of the VMSP runtime node to ensure all objects align with the updated template.

It is recommended to contact to Broadcom customer support if any issue or unclear on above steps