How to manually update kubeconfig/cluster configuration

Products

• VMware Telco Cloud Automation

Issue/Introduction

• K8S VIMs are disconnected in TCA-M, or the status of a cluster is disconnected (indicated with a red dot) in TCA-CP Appliance Management portal (9443).
• In TCA UI under partner system, Harbor is registered and connected Kubernetes cluster as VIM. In cluster Harbor add-on gets added successfully. But in partner system still it shows "Initiated", it should show "Enable".
• CNF installations fail.
• Attempts to create a new cluster results in the message: "Token has expired. Please login again."

Environment

2.x
3.x

Cause

• The cluster config stored inside TCA-DB for the clusters has expired
• The cluster certificates are renewed, but not synced with the TCA appliances.

Resolution

Manually update the kube-config/cluster configuration

1. SSH to the disconnected k8s cluster's VIP as the capv user.
2. View the contents of the '/etc/kubernetes/admin.conf' file.
   
   cat /etc/kubernetes/admin.conf
 
   

   apiVersion: v1
clusters:
- cluster:
    certificate-authority-data: [base64 CA data]
    server: https://<cluster VIP>:6443
  name: <cluster name>
contexts:
- context:
    cluster: <cluster name>
    user: kubernetes-admin
  name: kubernetes-admin@<cluster name>
current-context: kubernetes-admin@<cluster name>
kind: Config
preferences: {}
users:
- name: kubernetes-admin
  user:
    client-certificate-data: [base64 client cert data]
    client-key-data: [base64 client key data]
 

3. Copy the content of the 'admin.conf' file.
4. Go to the TCA-UI -> Virtual Infrastructure -> cluster.
5. Click on the 'Edit' button.
6. Update the Kubernetes Config with the admin.conf content obtained from step 3.
7. Click on 'Update'
   • This will set the VIM status to pending for about 30-40 seconds. It should return to the 'connected' status.
8. Navigate to https://<TCA-CP_IP>:9443 in a browser
9. Locate the cluster, edit and update the kube config file with the admin.conf content obtained from step 3.
10. Repeat steps 1 through 9 for any other affected clusters.
11. Restart the TCA Manager appliance
12. Restart the TCA CP Appliance
13. Verify the K8S VIMs status in TCA-M, or the status of the Workload Cluster in the TCA-CP Appliance Management portal (9443). They should be in a connected state.