Kubectl Commands Not Working After v1.20 Cluster Upgrade

Problem

After upgrading our cluster to v1.20, kubectl command returns connection refused message:

xxxxxxxxxx

The connection to the server localhost:8080 was refused - did you specify the right host or port?

Environment

• Platform9 Managed Kubernetes - v5.2 and Higher
• Cluster running on v1.20

Cause

This is an expected behavior from Kubernetes version v1.20

Resolution

• A KUBECONFIG environment variable will have to be explicitly sourced or the kubeconfig file needs to be specified via command line option --kubeconfig on the master nodes for kubectl to work.

Sample commands

xxxxxxxxxx

kubectl --kubeconfig=/path/to/kubeconfig/file get node

OR

export KUBECONFIG=/path/to/kubeconfig/file 

kubectl get nodes

​It is also recommended to use the kubeconfig assigned to users instead of using the file at /etc/pf9/kube.d/kubeconfigs/admin.yaml since it has blanket elevated permissions.