Getting Started with k0s and libvirt

This example shows how to create a single-node k0s cluster in a local VM using Terraform and libvirt. The blueprint-operator may then be installed.

Prerequisites

Along with blueprint CLI, you will also need the following tools installed:

k0sctl - required for installing a k0s distribution
terraform - for creating VMs
libvirt - required for running the local VM
terraform-provider-libvirt - required for provisioning libvirt VMs through terraform

You will also need an SSH key for authentication with the created VM.

Create virtual machines

Creating virtual machines can be done using the example Terraform scripts.

After copying the example TF scripts to your local machine, you can create the VMs with the following steps:

Create a terraform.tfvars file with content similar to:

cluster_name = "blueprint-cluster"
cores = 2
mem_size = "2048"
disk_size = 20
user_account = "user"
ssh_public_key = "ssh-rsa AAAEXAMPLE user@example.com"
ssh_key_path = "/home/user/.ssh/id_rsa"

terraform init
terraform plan
terraform apply
terraform output --raw k0s_cluster > cluster.yaml

To get detailed information about the created VMs, you can use virsh:
virsh machine list

Install Blueprint Operator on `k0s`

Verify or edit the blueprint created in the previous step, cluster.yaml
Create the cluster:

bctl apply -f cluster.yaml

Note: bctl apply adds kube config context to default location and sets it as the current context

Update the cluster by modifying cluster.yaml and then running:

bctl update -f cluster.yaml

Monitor the status of the cluster’s Kubernetes pods with:

watch -n 1 kubectl get pods --all-namespaces

Accessing the cluster

The example app addon can now be accessed through the http://<vm-node-ip>:6443 URL.

Cleanup

Delete the cluster:

bctl reset -f cluster.yaml

Delete virtual machines by changing to the example TF folder and running:

terraform destroy

Getting Started with an existing cluster Getting Started with k0s and Terraform