training-minikube

This project contains files to extend the original container image with Jupyter notebooks, including a bash kernel, Systemd service file and a Go development environment.

Minikube is the testbed to test Kubernetes, Jupyter is the standard for using notebooks running complex scenarios.

Goals

The intended use is for trainings, where this image is run in a privileged container.

k3s

Apply the secret to login first

kubectl create -f secret-training.yaml

You can obtain the token to login with

kubectl get secret training -o jsonpath='{.data.token}' | base64 --decode

Finally run the very privileged container on the first host

kubectl create -f k3s-training.yaml

Access

The container is deployed on host k3d-host-cluster-server-0. Adapt if your cluster has different names. You can now open http://k3d-host-cluster-server-0:8889/lab in a browser.

Change the name of the URI according to your setup. If you run k3s on the same node as the browser just use http://localhost:8889/lab

The port can be configured in the k3s-training.yaml. Just change JUPYTERLAB_PORT to your needs.

Dockerfile

You can bake your own images and add additional tools. This might be useful especially for air gapped environments.

If your pod has internet access you can additionally add tools add runtime from inside

Have fun!

Enjoy and give feedback! Especially if you make this run on different kinds of Kubernetes