This topic provides you the procedure to increase the existing worker node count by one to accomodate a newly added VMware Telco Cloud Service Assurance services in TKG, Azure, TCA, and VM Based.

Add a Node to Kubernetes Cluster Woker Node in TKG

  1. Get existing worker node count using the following command.
    export KUBECONFIG=/root/.kube/<clustername>;kubectl get nodes|grep -v NAME|grep -iv control|wc -l
  2. Scale worker node using the following command.
    tanzu cluster scale <clustername>  --worker-machine-count <current count + 1>

Add a Node to Kubernetes Cluster Woker Node in Azure

az aks scale --resource-group <your Resource-Group> --name <name of cluster> --node-count <number_of_vms> --nodepool-name <name of nodepool>
For example: 
az aks scale --resource-group rg-vmw-us-west --name tcsa-cluster --node-count 20 -nodepool-name nodepool

Add a Node to Kubernetes Cluster Woker Node TCA

  1. Launch VMware Telco Cloud Automation, navigate to Infrastructure > CaaS Infrastructure and click the Workload Cluster link that you want scale Worker Nodes.
  2. Click Node Pools tab.
  3. Click the vertical ellipsis (⋮) of the Worker Node pool.
  4. Click Edit.

    The Node Pool Details window appears in a while.

  5. In the Node Pool Details window, scroll down to the Number of replicas field and increase one extra node from the existing number of nodes.
  6. Click Apply.
  7. In the Workload Cluster: Edit Node Pool window, click Next.
  8. Click Update Node Pool.

    After the Worker Node Pool is increased, you will get a Ready message in the Node Pools section of the Workload Cluster window as shown in the following image.

Add a Node to Kubernetes Cluster Woker Node VMBased Deployment

Create a VM in the vCentre with static IP address in the same subnet where the existing cluster VMs are running and make it as a worker node by performing the following steps.
  1. Use the same vars.yaml file which was used during deployment and add a <new VM IP> to $HOME/k8s-installer/scripts/ansible/vars.yml file inside the Deployment Container as in the following file.
    worker_node_ips: # The list of IP addresses of your VMs. This should be a YML list.
    - <IP1> #By default IP1 to IPn will be available.
    - <IPn>
    - <new VM IP>
  2. Prepare for scale out and generate the inventory file ($HOME/k8s-installer/inventory/<ClusterName>/host.yml) by running the following command inside the Deployment Container.
    cd $HOME/k8s-installer
export ANSIBLE_CONFIG=$HOME/k8s-installer/scripts/ansible/ansible.cfg LANG=en_US.UTF-8
ansible-playbook scripts/ansible/prepare.yml -e @scripts/ansible/vars.yml --skip-tags check-vm-resources --become
  3. Scale out the Kubernetes cluster using the generated inventory file ($HOME/k8s-installer/inventory/<ClusterName>/host.yml) by running the following command inside the Deployment Container.
    cd $HOME/k8s-installer/
ansible-playbook -i inventory/<your-cluster-name>/hosts.yml scripts/ansible/scale_k8s.yml -u <your-SSH-username> --become -e @scripts/ansible/vars.yml -e @scripts/ansible/internal_vars.yml --limit="localhost,nodex"
    Note: The --limit argument must be used to limit the scale operation to the new nodes added to the inventory file. The node names can be obtained from the generated inventory file.
    • For example, if you originally had three nodes and you added a fourth node, then the node name to be used for the scale operation must be node4.
    • Localhost is mandatory for the --limit argument.

      If you are scaling multiple nodes, then --limit must be comma separated --limit=localhost, nodex, nodey.

  4. Verify the scaled cluster inside the control node.
    kubectl get nodes (new node will show "READY" status)