New Worker/Control Plane Nodes Fail to Deploy Due to NSX Network Realization Issue

Products

• VMware vSphere Kubernetes Service

Issue/Introduction

New worker or control plane nodes are not deploying. Deployment attempts fail with the following errors:

• InfrastructureReady: network interface "eth0" error: network interface is not ready: NetworkNotRealized - Cannot realize network

• NodeHealthy: Waiting for VSphereMachine to report spec.providerID

Environment

• VKS
• vSphere with Tanzu

Resolution

The issue resolves by restarting the following deployments on the Supervisor cluster to restore NSX and Cluster API functionality:

Deployment	Namespace
capi-controller-manager	svc-tkg-domain-c####
capi-kubeadm-bootstrap-controller-manager	svc-tkg-domain-c####
capv-controller-manager	svc-tkg-domain-c####
vmware-system-vmop-controller-manager	vmware-system-vmop
nsx-ncp	vmware-system-nsx

Command: kubectl rollout restart deployment <deployment_name> -n <namespace>

After restarting these components, the new nodes are successfully provisioned and reach the Running state.