Adding Servers

1. Create a new server that meets the system requirements
2. On your existing server run the following command:

unbind add-node

This will output a command that looks like:

curl -sfL https://get.k3s.io | INSTALL_K3S_VERSION=v1.33.1+k3s1 K3S_URL=https://<YOUR_SERVER_IP>:6443 K3S_TOKEN=<SECRET_JOIN_TOKEN> sh -

3. Copy the command and run it on the new server.

Verifying the Node joined.

Run the following command on your primary server to check if the node joined successfully:

kubectl get nodes

Removing a node

1. Cordon the node you intend to remove, this will prevent new workloads from being scheduled to it.

kubectl cordon <node-name>

2. Drain the node, this will gracefully move all workloads to other nodes.

kubectl drain <node-name> --ignore-daemonsets --delete-local-data

3. Delete the node.

kubectl delete node <node-name>

4. Verify the node is removed.

kubectl get nodes

You are free to repurpose or delete the server as you see fit.

(Advanced) High Availability

For a high availability k3s cluster, generally you will need to configure the following:

• An odd number of control plane servers (at least 3).
• Control planes need to use etcd, for k3s this is a matter of adding --cluster-init to /etc/systemd/system/k3s.service and restarting the service with sudo systemctl daemon-reload && sudo systemctl restart k3s
• It’s ideal to configure an external load balancer, many cloud providers offer load balancers as a service.
• It’s ideal to not schedule normal workloads on the control plane servers, and instead use a separate set of servers for that.
  • You can use taints to prevent workloads from being scheduled to the control plane servers.
• If using longhorn (Unbind’s default storage class) - it’s ideal to increase the number of replicas to 3 or more to prevent data loss in the event of a total system failure.