Skip to content

Creating a cluster with Ansible Playbook#

Using Ansible and the k0s-ansible playbook, you can install a multi-node Kubernetes Cluster in a couple of minutes. Ansible is a popular infrastructure as code tool which helps you automate tasks to achieve the desired state in a system.

This guide shows how you can install k0s on local virtual machines. In this guide, the following tools are used:

  • multipass, a lightweight VM manager that uses KVM on Linux, Hyper-V on Windows, and hypervisor.framework on macOS (installation guide).
  • ansible, a popular infrastructure as code tool (installation guide).
  • and of course kubectl on your local machine (Installation guide).

Before following this tutorial, you should have a general understanding of Ansible. A great way to start is the official Ansible User Guide.

Please note: k0s users created k0s-ansible. Please send your feedback, bug reports, and pull requests to

Without further ado, let's jump right in.

Download k0s-ansible#

On your local machine clone the k0s-ansible repository:

$ git clone
$ cd k0s-ansible

Create virtual machines#

For this tutorial, multipass was used. However, there is no interdependence. This playbook should also work with VMs created in alternative ways or Raspberry Pis.

Next, create a couple of virtual machines. For the automation to work, each instance must have passwordless SSH access. To achieve this, we provision each instance with a cloud-init manifest that imports your current users' public SSH key and into a user k0s. For your convenience, a bash script is included that does just that:

./tools/ 7 ◀️ this creates 7 virtual machines

$ ./tools/ 7
Create cloud-init to import ssh key...
[1/7] Creating instance k0s-1 with multipass...
Launched: k0s-1
[2/7] Creating instance k0s-2 with multipass...
Launched: k0s-2
[3/7] Creating instance k0s-3 with multipass...
Launched: k0s-3
[4/7] Creating instance k0s-4 with multipass...
Launched: k0s-4
[5/7] Creating instance k0s-5 with multipass...
Launched: k0s-5
[6/7] Creating instance k0s-6 with multipass...
Launched: k0s-6
[7/7] Creating instance k0s-7 with multipass...
Launched: k0s-7
Name State IPv4 Image
k0s-1 Running Ubuntu 20.04 LTS
k0s-2 Running Ubuntu 20.04 LTS
k0s-3 Running Ubuntu 20.04 LTS
k0s-4 Running Ubuntu 20.04 LTS
k0s-5 Running Ubuntu 20.04 LTS
k0s-6 Running Ubuntu 20.04 LTS
k0s-7 Running Ubuntu 20.04 LTS

Create Ansible inventory#

After that, we create our inventory directory by copying the sample:

$ cp -rfp inventory/sample inventory/multipass

Now we need to create our inventory. The before built virtual machines need to be assigned to the different host groups required by the playbook's logic.

  • initial_controller = must contain a single node that creates the worker and controller tokens needed by the other nodes.
  • controller = can contain nodes that, together with the host from initial_controller form a highly available isolated control plane.
  • worker = must contain at least one node so that we can deploy Kubernetes objects.

We could fill inventory/multipass/inventory.yml by hand with the metadata provided by multipass list, but since we are lazy and want to automate as much as possible, we can use the included Python script

To automatically fill our inventory run:

$ ./tools/
Designate first three instances as control plane
Created Ansible Inventory at: /Users/dev/k0s-ansible/tools/inventory.yml
$ cp tools/inventory.yml inventory/multipass/inventory.yml

Now inventory/multipass/inventory.yml should look like this (Of course, your IP addresses might differ):

    ansible_user: k0s

Test the connection to the virtual machines#

To test the connection to your hosts just run:

$ ansible -i inventory/multipass/inventory.yml -m ping
k0s-4 | SUCCESS => {
    "ansible_facts": {
        "discovered_interpreter_python": "/usr/bin/python3"
    "changed": false,
    "ping": "pong"

If all is green and successful, you can proceed.

Provision the cluster with Ansible#

Finally, we can start provisioning the cluster. Applying the playbook, k0s will get downloaded and set up on all nodes, tokens will get exchanged, and a kubeconfig will get dumped to your local deployment environment.

$ ansible-playbook site.yml -i inventory/multipass/inventory.yml
TASK [k0s/initial_controller : print kubeconfig command] *******************************************************
Tuesday 22 December 2020  17:43:20 +0100 (0:00:00.257)       0:00:41.287 ******
ok: [k0s-1] => {
    "msg": "To use Cluster: export KUBECONFIG=/Users/dev/k0s-ansible/inventory/multipass/artifacts/k0s-kubeconfig.yml"
PLAY RECAP *****************************************************************************************************
k0s-1                      : ok=21   changed=11   unreachable=0    failed=0    skipped=1    rescued=0    ignored=0
k0s-2                      : ok=10   changed=5    unreachable=0    failed=0    skipped=1    rescued=0    ignored=0
k0s-3                      : ok=10   changed=5    unreachable=0    failed=0    skipped=1    rescued=0    ignored=0
k0s-4                      : ok=9    changed=5    unreachable=0    failed=0    skipped=1    rescued=0    ignored=0
k0s-5                      : ok=9    changed=5    unreachable=0    failed=0    skipped=1    rescued=0    ignored=0
k0s-6                      : ok=9    changed=5    unreachable=0    failed=0    skipped=1    rescued=0    ignored=0
k0s-7                      : ok=9    changed=5    unreachable=0    failed=0    skipped=1    rescued=0    ignored=0

Tuesday 22 December 2020  17:43:36 +0100 (0:00:01.204)       0:00:57.478 ******
prereq : Install apt packages -------------------------------------------------------------------------- 22.70s
k0s/controller : Wait for k8s apiserver ----------------------------------------------------------------- 4.30s
k0s/initial_controller : Create worker join token ------------------------------------------------------- 3.38s
k0s/initial_controller : Wait for k8s apiserver --------------------------------------------------------- 3.36s
download : Download k0s binary k0s-v0.9.0-rc1-amd64 ----------------------------------------------------- 3.11s
Gathering Facts ----------------------------------------------------------------------------------------- 2.85s
Gathering Facts ----------------------------------------------------------------------------------------- 1.95s
prereq : Create k0s Directories ------------------------------------------------------------------------- 1.53s
k0s/worker : Enable and check k0s service --------------------------------------------------------------- 1.20s
prereq : Write the k0s config file ---------------------------------------------------------------------- 1.09s
k0s/initial_controller : Enable and check k0s service --------------------------------------------------- 0.94s
k0s/controller : Enable and check k0s service ----------------------------------------------------------- 0.73s
Gathering Facts ----------------------------------------------------------------------------------------- 0.71s
Gathering Facts ----------------------------------------------------------------------------------------- 0.66s
Gathering Facts ----------------------------------------------------------------------------------------- 0.64s
k0s/worker : Write the k0s token file on worker --------------------------------------------------------- 0.64s
k0s/worker : Copy k0s service file ---------------------------------------------------------------------- 0.53s
k0s/controller : Write the k0s token file on controller ------------------------------------------------- 0.41s
k0s/controller : Copy k0s service file ------------------------------------------------------------------ 0.40s
k0s/initial_controller : Copy k0s service file ---------------------------------------------------------- 0.36s

Use the cluster with kubectl#

While the playbook ran, a kubeconfig got copied to your local machine. You can use it to get simple access to your new Kubernetes cluster:

$ export KUBECONFIG=/Users/dev/k0s-ansible/inventory/multipass/artifacts/k0s-kubeconfig.yml
$ kubectl cluster-info
Kubernetes control plane is running at
CoreDNS is running at
Metrics-server is running at

$ kubectl get nodes -o wide
k0s-4   Ready      <none>   21s   v1.20.1-k0s1   <none>        Ubuntu 20.04.1 LTS   5.4.0-54-generic   containerd://1.4.3
k0s-5   Ready      <none>   21s   v1.20.1-k0s1   <none>        Ubuntu 20.04.1 LTS   5.4.0-54-generic   containerd://1.4.3
k0s-6   NotReady   <none>   21s   v1.20.1-k0s1   <none>        Ubuntu 20.04.1 LTS   5.4.0-54-generic   containerd://1.4.3
k0s-7   NotReady   <none>   21s   v1.20.1-k0s1   <none>        Ubuntu 20.04.1 LTS   5.4.0-54-generic   containerd://1.4.3

⬆️ Of course, the first three control plane nodes won't show up here because the control plane is fully isolated. You can check on the distributed etcd cluster by running this ad-hoc command (or ssh'ing directly into a controller node):

$ ansible k0s-1 -a "k0s etcd member-list -c /etc/k0s/k0s.yaml" -i inventory/multipass/inventory.yml | tail -1 | jq
  "level": "info",
  "members": {
    "k0s-1": "",
    "k0s-2": "",
    "k0s-3": ""
  "msg": "done",
  "time": "2020-12-23T00:21:22+01:00"

After a while, all worker nodes become Ready. Your cluster is now waiting to get used. We can test by creating a simple nginx deployment.

$ kubectl create deployment nginx --replicas=5
deployment.apps/nginx created

$ kubectl expose deployment nginx --target-port=80 --port=8100
service/nginx exposed

$ kubectl run hello-k0s --rm -it --restart=Never --command -- wget -qO- nginx:8100
<!DOCTYPE html>
<title>Welcome to nginx on Debian!</title>
pod "hello-k0s" deleted