VKS Cluster status False after upgrading to VKS 3.6 and VCF 9.1

Products

• VMware vSphere Kubernetes Service

Issue/Introduction

After upgrading to VMware Kubernetes Service (VKS) 3.6 and VCF 9.1, clusters may report a "False" status for the TopologyReconciled condition. This occurs when a single-zone namespace has been migrated to a multi-zone configuration (associated with multiple zones) but the cluster lacks explicit failureDomain definitions while using a StorageClass with Immediate binding mode.

Symptoms include the following error message in the cluster status:

error computing the desired state of the Cluster topology: failed to apply patches: failed to generate patches for patch "default": failed to call extension handler "generate-patches.runtime-extension": got failure response.

Environment

• VMware Cloud Foundation (VCF) 9.1

• VMware Kubernetes Service (VKS) 3.6

Cause

VKS 3.6 and VCF 9.1 require clusters to utilize a StorageClass with WaitForFirstConsumer binding mode if no failureDomain is explicitly specified in the cluster configuration when the namespace spans multiple zones.

Resolution

To restore the cluster status, perform one of the following manual updates to the Cluster specification:

Explicitly define Failure Domains: Update the cluster YAML to include the failureDomain for each Node Pool.

Switch Storage Binding Mode: Update the storageClass in the cluster specification to a version that utilizes WaitForFirstConsumer binding (identifiable by the -latebinding suffix).

Note: Applying either of these changes will trigger a rolling update of the affected node pools.

Additional Information

• VKS/Tanzu Support Definitions

• VCF Content Guidelines