Adding and Removing Nodes in Cluster

Guide to Adding and Removing Nodes from an Existing Kubernetes Cluster

Updated automation guide: For the current OpenG2P production add-node workflow (bundled scripts, step-by-step), see Add Node. Optional AWS EC2 provisioning is documented separately at AWS Add-Node Provisioning (optional).

Adding Nodes to Cluster

Below are the steps to add new nodes to an existing Kubernetes cluster.

• From kubernetes/rke2 directory, use the rke2-server.conf.subsequent.template or rke2-agent.conf.template based on whether the new node is control plane node or worker node. Copy this file to /etc/rancher/rke2/config.yaml in the new node.

• Configure the config.yaml with relevant values.

• Run the following to set the RKE2 version. Make sure this version is the same across all the nodes. (Refer to RKE2 Releases. Use rke2 --version on an existing node to get the version.)

  export INSTALL_RKE2_VERSION="v1.35.8+rke2r1"

• Run the below command to download RKE2.

  curl -sfL https://get.rke2.io | sh -

• Run the below commands to start RKE2 node.

  systemctl enable rke2-server
  systemctl start rke2-server

Deleting Nodes from Cluster

Follow the steps below to remove a node from the Kubernetes cluster.

• Before you delete the node, make sure the PodDisruptionBudget is set to "0" on deleting node. Click here for more information.

• To drain the node from the cluster, run the command below.

  kubectl drain <nodename> --ignore-daemonsets --delete-emptydir-data

• Once done draining the node, run the command below to delete it from the cluster.

  kubectl delete node <nodename>

• After deleting the node, check your Kubernetes cluster to ensure the node has been deleted.

• Make sure the node IP is removed from the LoadBalancer/NGINX to avoid intermittent issues in the environment.