Skip to content

Worker Node Setup

Worker nodes run your actual workloads (pods). They join the cluster by connecting through the Nginx load balancer, which routes to whichever master is available.

Prerequisites

  • At least one master node is up and running
  • You have the node token from Master 1 (/var/lib/rancher/k3s/server/node-token)
  • The Nginx load balancer is reachable at 192.168.0.156:6443

Joining a Worker Node

Run the following command on each worker machine. Replace <token-goes-here> with the token from Master 1.

curl -sfL https://get.k3s.io | K3S_URL=https://192.168.0.156:6443 K3S_TOKEN=<token-goes-here> sh -

This single command:

  1. Downloads and installs the k3s agent
  2. Registers the worker with the cluster via the load balancer
  3. Starts the agent as a systemd service

Worker 1

Worker 1
curl -sfL https://get.k3s.io | K3S_URL=https://192.168.0.156:6443 K3S_TOKEN=<token-goes-here> sh -

Worker 2

Worker 2
curl -sfL https://get.k3s.io | K3S_URL=https://192.168.0.156:6443 K3S_TOKEN=<token-goes-here> sh -

Worker 3

Worker 3
curl -sfL https://get.k3s.io | K3S_URL=https://192.168.0.156:6443 K3S_TOKEN=<token-goes-here> sh -

Verify All Nodes

From any master node, confirm all workers appear in the node list:

sudo k3s kubectl get nodes -o wide

You should see all master and worker nodes listed with status Ready.

Adding More Workers

To scale the cluster, simply run the join command on any additional machines. No changes to the master nodes are needed.

Next Step

Continue to MetalLB Setup to configure bare-metal LoadBalancer IP assignment.