Skip to main content
Version: 3.28 (latest)

K3s multi-node install

Big picture

This tutorial gets you a multi node K3s cluster with Calico in approximately 10 minutes.


K3s is a lightweight implementation of Kubernetes packaged as a single binary.

The geeky details of what you get:


Before you begin

  • Make sure you have a linux host that meets the following requirements
    • x86-64 processor
    • 1CPU
    • 1GB Ram
    • 10GB free disk space
    • Ubuntu 18.04 (amd64), Ubuntu 20.04 (amd64)

K3s supports ARM processors too, this tutorial was tested against x86-64 processor environment. For more detail please visit this link.

How to

Initializing control plane instance

K3s installation script can be modified by environment variables. Here you are providing some extra arguments to disable flannel, disable k3s default network policy and change the pod ip CIDR.


Full list of arguments can be viewed at this link.

curl -sfL | INSTALL_K3S_EXEC="--flannel-backend=none --disable-network-policy --cluster-cidr=" sh -

If is already in use within your network you must select a different pod network CIDR by replacing in the above command.

Enable remote access to your K3s instance

To set up remote access to your cluster first ensure you have installed kubectl on your system.


If you are not sure how to install kubectl in your OS visit this link.

K3s stores a kubeconfig file in your server at /etc/rancher/k3s/k3s.yaml, copy all the content of k3s.yaml from your server into ~/.kube/config on the system that you like to have remote access to the cluster.

Add extra nodes to K3s cluster

To add additional nodes to your cluster you need two piece of information.

  • K3S_URL which is going to be your main node ip address.
  • K3S_TOKEN which is stored in /var/lib/rancher/k3s/server/node-token file in main Node (Step 1). Execute following command in your node instance and join it to the cluster.

Remember to change serverip and mytoken.

curl -sfL | K3S_URL=https://serverip:6443 K3S_TOKEN=mytoken sh -

Install Calico

Install the Calico operator and custom resource definitions.

kubectl create -f

Due to the large size of the CRD bundle, kubectl apply might exceed request limits. Instead, use kubectl create or kubectl replace.

Install Calico by creating the necessary custom resource. For more information on configuration options available in this manifest, see the installation reference.

kubectl create -f

Before creating this manifest, read its contents and make sure its settings are correct for your environment. For example, you may need to change the default IP pool CIDR to match your pod network CIDR.

Check the installation

  1. Confirm that all of the pods are running using the following command.
NAMESPACE         NAME                                      READY   STATUS    RESTARTS   AGE
tigera-operator tigera-operator-c9cf5b94d-gj9qp 1/1 Running 0 107s
calico-system calico-typha-7dcd87597-npqsf 1/1 Running 0 88s
calico-system calico-node-rdwwz 1/1 Running 0 88s
kube-system local-path-provisioner-6d59f47c7-4q8l2 1/1 Running 0 2m14s
kube-system metrics-server-7566d596c8-xf66d 1/1 Running 0 2m14s
kube-system coredns-8655855d6-wfdbm 1/1 Running 0 2m14s
calico-system calico-kube-controllers-89df8c6f8-7hxc5 1/1 Running 0 87s
  1. Confirm that you now have two nodes in your cluster with the following command.

    kubectl get nodes -o wide

    It should return something like the following.

    k3s-master Ready master 40m v1.18.2+k3s1 <none> Ubuntu 18.04.3 LTS 4.15.0-101-generic containerd://1.3.3-k3s2
    k3s-node1 Ready <none> 30m v1.18.2+k3s1 <none> Ubuntu 18.04.3 LTS 4.15.0-101-generic containerd://1.3.3-k3s2

Congratulations! You now have a multi node K3s cluster equipped with Calico and Traefik.

Next steps