fcrespel/k8s-vagrant-node

Kubernetes Vagrant node (master/worker) for local testing with CRI-O, Calico, WireGuard, Ingress Nginx and OpenEBS

Kubernetes Vagrant node

This repository contains a Vagrant VM to run a Kubernetes node (master or worker) for local testing/debugging, using:

• CRI-O as the container runtime
• Calico as the network plugin
• WireGuard to encrypt communication between nodes
• Ingress Nginx to expose ingress resources
• OpenEBS to support local persistent volumes

Prerequisites

Make sure you have VirtualBox and Vagrant installed before starting.

You may clone this repository mutliple times in different folders to build a cluster with multiple VMs (e.g. 1 master and 2 workers).

Configuration

First, configure the Vagrantfile:

• Update the private_network IP address (e.g. 192.168.56.101 for master, 192.168.56.102 for worker).
• Adjust vb.memory depending on your available RAM and node role (at least 2048 MB for a master, less for a worker).

Edit the provision.conf.sh file:

• Update K8S_MASTER_IP if you changed it in the master Vagrantfile
• Set K8S_NODE_ROLE to master or worker
• Set K8S_NODE_NAME to an arbitrary name (e.g. node1, node2, ...)

For a worker node, you will need to obtain a token and certificate hash from the master first. This information is displayed at the end of master creation, if necessary you can get it again with:

kubeadm token create

openssl x509 -pubkey -in /etc/kubernetes/pki/ca.crt | openssl rsa -pubin -outform der 2>/dev/null | openssl dgst -sha256 -hex | sed 's/^.* //'

Then, update the worker provision.conf.sh:

• Set K8S_TOKEN to the join token
• Set K8S_CERTHASH to the certificate hash

Running

Start the VM by running vagrant up in the repository directory. The provisioning script will take some time to install and configure the node.

When done or if an error occurs, you should restart the VM with vagrant reload to make sure the updated kernel and modules are properly loaded.

After that, you may get a shell in the VM with vagrant ssh, and run commands such as the following:

# Switch to root
sudo su -

# Check pod/deployment status
kubectl get all -A

# Check Calico node status
calicoctl node status

# Check WireGuard status
wg

# Check containers running locally
crictl ps

If everything is OK, you may deploy pods with kubectl or helm, or configure Calico resources with calicoctl. You may also use k9s to get an interactive text-based interface; kubectx and kubens utilities are also available.

Note that Ingress Nginx is preinstalled and exposed on node ports 30080 (HTTP) and 30443 (HTTPS).