Deploying a VKS guest cluster against NSX is failing with error, "unable to reconcile configmap's coredns info: unable to retrieve kubeadm configmap from the guest cluster failed to get server groups: get "https:/<clusterControlPlane>:6443/api?timeout=10"

Products

• VMware vSphere Kubernetes Service

Issue/Introduction

• A new cluster is configured against a Supervisor but the worker nodes are not created.
• Running a describe against the new cluster shows the error, "Error, "unable to reconcile configmap's coredns info: unable to retrieve kubeadm configmap from the guest cluster failed to get server groups: get "https:/<clusterControlPlane>:6443/api?timeout=10s: net/http request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)"
• NSX Traceflow from Supervisor eth1 to the new cluster node succeeds.
• ping -s 1400 from supervisor to cluster node succeeds, but ping -s 1500 fails.  
• MTU consistency check in NSX is successful.

Cause

Insufficient MTU configured on the physical layer to support traffic on NSX overlay.

Resolution

• Troubleshoot and resolve physical configuration MTU for NSX traffic until MTU 1500 pings are successful.

Additional Information

Troubleshooting MTU issues and fragmentation in NSX fabric

Networking Requirements for downloading VKS Supervisor Services