Multi-nodes cluster

In this part, you will create a multi-nodes cluster.

Create the virtual machines

First, run the following command to launch VMs named k0s-1, k0s-2 and k0s-3.

multipass launch -n k0s-1 --disk 10G --cpus 2 --memory 2G
multipass launch -n k0s-2 --disk 10G --cpus 2 --memory 2G
multipass launch -n k0s-3 --disk 10G --cpus 2 --memory 2G

Next list the VMs and make sure they are running fine:

multipass list

You should get an output similar to the following one (the IP addresses you get would be different though).

Name                    State             IPv4             Image
k0s-1                   Running           192.168.64.27    Ubuntu 24.04 LTS
k0s-2                   Running           192.168.64.25    Ubuntu 24.04 LTS
k0s-3                   Running           192.168.64.26    Ubuntu 24.04 LTS

Download k0s binary in each VMs 

for i in 1 2 3; do
  multipass exec k0s-$i -- /bin/sh -c "curl -sSf https://get.k0s.sh | sudo sh"
done

Init the cluster

First, run a shell in k0s-1.

multipass shell k0s-1

Next, install k0s controller and start it.

sudo k0s install controller
sudo k0s start
We don’t use the --single option as this instance will only act as a control plane Node.

You can now exit the shell from k0s-1.

Adding worker Nodes

In order to add worker Nodes, we need to generate a token from the control plane and use this token to install k0s on the worker Nodes.

First, generate the token as follows. This will create the file worker_token in the current folder.

multipass exec k0s-1 -- sudo k0s token create --role=worker > ./worker_token
The worker role specified in the command indicates that the token will be used to add a worker (default value). We could also use a controller role to add additional controllers in the cluster.

Next, copy that token into k0s-2 and k0s-3.

multipass transfer ./worker_token k0s-2:/tmp/worker_token
multipass transfer ./worker_token k0s-3:/tmp/worker_token

Next, install k0s onto k0s-2 and k0s-3 VMs.

multipass exec k0s-2 -- sudo k0s install worker --token-file /tmp/worker_token
multipass exec k0s-3 -- sudo k0s install worker --token-file /tmp/worker_token

Then, start k0s on both worker Nodes:

multipass exec k0s-2 -- sudo k0s start
multipass exec k0s-3 -- sudo k0s start

Get a kubeconfig file 

multipass exec k0s-1 -- sudo cat /var/lib/k0s/pki/admin.conf > k0s.kubeconfig

Next, replace localhost with the actual IP address of the Multipass VM.

K0S1_IP=$(multipass info k0s-1 | grep IP | awk '{print $2}')
sed -i'' "s/localhost/$K0S1_IP/" k0s.kubeconfig

Then, configure your local kubectl, so it uses this modified kubeconfig file.

export KUBECONFIG=$PWD/k0s.kubeconfig

Listing the cluster’s Nodes you should get a list similar to the one below.

$ kubectl get nodes
NAME    STATUS   ROLES    AGE   VERSION
k0s-2   Ready    <none>   92s   v1.33.4+k0s
k0s-3   Ready    <none>   92s   v1.33.4+k0s
As k0s-1 was configured as a control plane only Node, it does not appear in this list which only shows the worker Nodes.

Cleanup

You can now delete the VMs used in this section.

multipass delete -p k0s-1 k0s-2 k0s-3

In the next section, you will use k0sctl, a k0s’s companion tool that simplifies the management of multi-nodes clusters.