VMware Telco Cloud Automation is integrated with VMware Tanzu Kubernetes Grid. To use the latest version 1.6.1 of Tanzu Kubernetes Grid, you must upgrade the Tanzu Kubernetes management first and then workload clusters.
For information about the Kubernetes upgrade compatibility for the clusters, see Upgrade Kubernetes Version.
- Upgrade the Tanzu Kubernetes management clusters.
- Download the new version of the Kubernetes Cluster template that you use to upgrade the Tanzu Kubernetes cluster. For instructions to download Kubernetes Cluster templates, see the Download Photon BYOI Templates for VMware Tanzu Kubernetes Grid section in the VMware Telco Cloud Automation 2.0 Release Notes.
- Import the Kubernetes Cluster template into vSphere where the Tanzu Kubernetes cluster is located.
- Upgrade the Tanzu Kubernetes cluster. For instructions, see Upgrade the Kubernetes Version for a Cluster Instance.
- (Optional) If you want to make any Kubernetes Cluster configuration changes, edit the template. For instructions, see Upgrade the Kubernetes version for Cluster Template.
- Repeat Steps 1a-1d to upgrade each Tanzu Kubernetes cluster.
- Upgrade the Tanzu Kubernetes workload clusters as follows:
- Deploy a new Tanzu Kubernetes workload cluster.
- Do the following on the old Tanzu Kubernetes workload cluster you want to upgrade:
Identify the Distributed Unit (DU) that you want to upgrade.
Acquire a soft lock on the DU.
Note: The soft lock stops accepting new UE sessions and hands off the existing UE sessions to any other active DU.
Collect the specification details of the node where the DU is hosted.
Terminate the DU Network Function (NF).
Note: Do not delete the NF instance.
- Create a new DU node pool on the new workload cluster with the same specification collected in Step 2b.
- Instantiate the NF instance in the new DU node pool.
- Change the cloud specification from the old DU node pool to the new DU node pool.
- Repeat Steps 2a-2e for all workload clusters.
VMware Tanzu Kubernetes Clusters are upgraded successfully. All pods in the cluster are up and running.