Adding a Node

This guide walks through adding a new physical machine to the cluster as a worker node.

Prerequisites

• A machine with Ubuntu Server installed
• Network connectivity to the campus LAN
• An available static IP in the 192.168.136.x range
• SSH access to the new machine and to nst-n1

Step 1: Install Ubuntu Server

Install Ubuntu Server on the new machine. During installation:

• Choose a hostname following the pattern: nst-n6, nst-n7, etc.
• Set the default boot target to multi-user.target (no GUI)
• Create a user account

Step 2: Configure static IP

Find the network interface and MAC address:

ip link show

Create a Netplan configuration:

sudo nano /etc/netplan/01-netcfg.yaml

network:
  ethernets:
    enp2s0:
      addresses:
        - 192.168.136.150/20    # use the next available IP
      routes:
        - to: default
          via: 192.168.128.1
      match:
        macaddress: XX:XX:XX:XX:XX:XX   # your machine's MAC
      nameservers:
        addresses: [8.8.8.8, 8.8.4.4]
      set-name: enp2s0
  version: 2

Apply:

sudo netplan apply

Verify connectivity:

ping 192.168.136.145   # should reach nst-n1
ping 8.8.8.8           # should reach the internet

Step 3: Set the hostname

sudo hostnamectl set-hostname nst-n6

Edit /etc/hosts to include the new hostname:

echo "192.168.136.150 nst-n6" | sudo tee -a /etc/hosts

Step 4: Get the join token

On nst-n1:

sudo cat /var/lib/rancher/k3s/server/node-token

Copy the token.

Step 5: Install K3s agent

On the new node:

curl -sfL https://get.k3s.io | K3S_URL=https://192.168.136.145:6443 K3S_TOKEN=<TOKEN> sh -

This installs K3s in agent (worker) mode and joins the cluster.

Step 6: Verify the node joined

On nst-n1:

kubectl get nodes -o wide

The new node should appear as Ready within a minute.

Step 7: Label the node

kubectl label node nst-n6 role=compute --overwrite

This allows JupyterHub and other workloads with nodeSelector: { role: compute } to schedule pods on this node.

Step 8: Verify workloads can schedule

Check that system pods (like svclb-traefik) are running on the new node:

kubectl get pods -A -o wide | grep nst-n6

You should see at least svclb-traefik pods.

Step 9: Update documentation

Add the new node to:

• The node inventory in these docs
• Any monitoring or alerting configurations
• The allnodes script if it uses a hardcoded node list

Post-setup checklist

• [ ] Static IP configured and tested
• [ ] Hostname set
• [ ] K3s agent installed and joined
• [ ] Node shows Ready in kubectl
• [ ] Node labeled role=compute
• [ ] svclb-traefik pod running on the node
• [ ] Documentation updated