Small deployment of Whole Kubernetes for development tests

This is a small tutorial/demo of how to deploy a whole K3S cluster in a single computer.

What is K3S

k3s is a Lightweight Kubernetes, it supports all the API calls/operators/etc that Kubernetes(k8s) support.

*BTW: k3s = k(8/2)s *

Why k3s

Is a single binary deployment, run on every machine (linux required), it supports boards with ARMV7+ (like raspberry pi), easy to install and easy to deploy.

Requirements

Windows x64 with WSL2 enable or a recent Linux OS, I'm using Ubuntu 20.04 Last version of (Docker)[] installed, I'm using 4.4.0 at this moment.

Installation

We are going to use k3d, it is a helpful tool that help us to deploy the cluster in docker with a single command line command.

BTW: k3d = k3s in Docker

Linux

You can check the script before run the commands at: https://github.com/rancher/k3d/blob/main/install.sh

With wget

wget -q -O - https://raw.githubusercontent.com/rancher/k3d/main/install.sh | bash

With curl

curl -s https://raw.githubusercontent.com/rancher/k3d/main/install.sh | bash

Windows

Make sure that you have WSL2 enable and Docker with WSL2 components.

With Chocolatey

1. Install Chocolatey https://chocolatey.org/install
2. run choco install k3d -y
3. run choco install kubernetes-cli

Manual Installation

1. Download k3d from the github https://github.com/rancher/k3d/releases/
2. Download kubectl from official site https://kubernetes.io/docs/tasks/tools/install-kubectl-windows/
3. Put those files on a folder
4. Add that folder to the system path

Starting the cluster

This command will simulate 3 computers, 2 as master nodes and 3 working nodes, additionally will forward any request to the ports 8888 and 8889 on the host system to the cluster.

Deploy the server

k3d cluster create localk8s --servers 1 --agents 3 --port "8888:80@loadbalancer" --port "8889:443@loadbalancer"

Check if it is running

kubectl get nodes --output wide

You should see something similar to:

kubectl get nodes --output wide
NAME                    STATUS   ROLES                       AGE   VERSION        INTERNAL-IP   EXTERNAL-IP   OS-IMAGE   KERNEL-VERSION                      CONTAINER-RUNTIME
k3d-localk8s-agent-0    Ready    <none>                      68m   v1.21.7+k3s1   172.18.0.7    <none>        Unknown    5.10.16.3-microsoft-standard-WSL2   containerd://1.4.12-k3s1
k3d-localk8s-agent-1    Ready    <none>                      68m   v1.21.7+k3s1   172.18.0.9    <none>        Unknown    5.10.16.3-microsoft-standard-WSL2   containerd://1.4.12-k3s1
k3d-localk8s-agent-2    Ready    <none>                      68m   v1.21.7+k3s1   172.18.0.5    <none>        Unknown    5.10.16.3-microsoft-standard-WSL2   containerd://1.4.12-k3s1
k3d-localk8s-server-0   Ready    control-plane,etcd,master   68m   v1.21.7+k3s1   172.18.0.3    <none>        Unknown    5.10.16.3-microsoft-standard-WSL2   containerd://1.4.12-k3s1

How to delete the cluster

k3d cluster delete localk8s

Ready for play

If something goes wrong just delete the cluster and deploy it again.