simple-kubernetes-deploy

A group of very simple Ansible playbooks to deploy the most basic Kubernetes Cluster

The main inspiration for these playbooks comes from this Digital Ocean article

Ultimately, the desire is to simplify the deployment of a basic Kubernetes cluster as much as possible without getting into any of the nasty details. The Kubernetes community have done fantastic work in the past few years to simplify kubeadm and other tools so if you want to pay them back, please consider becoming a K8s contributor

Getting Started

First make sure you align with the requirements set out below, then run the Ansible playbooks and finally test if the deployment was successful.

Requirements

There are three requirements to get started, make sure:

1. You have 4 nodes (Check out the Tested On Section to see tested OS')
   • 1 Node for Ansible Controller
   • 1 Node for Control Plane
   • 2 Nodes for Worker
2. You have Ansible installed on the Ansible Controller Ubuntu 20.04:

   sudo apt install ansible

3. Your Ansible Controller has root ssh key access to the Control Plane and Worker Nodes Ubuntu 20.04: Create the SSH Pub/Priv Keys:

   # On Ansible Controller:
   ssh-keygen -t ed25519

   Upload the ssh public key to your Github account. Then import them as root on your other nodes:

   # On Control Plane and Worker Nodes:
   sudo su # Enter user password at prompt
   
   ssh-import-id-gh <github-username>

hosts file

Add details about all the nodes you will deploy K8s on in the hosts file

Run the Ansible Playbook

To deploy the Playbook, simply run the following command:

ansible-playbook -i hosts k8s-deploy.yml

Assuming no errors came up, congratulations, you've deployed a Kubernetes Cluster!! 🎉 🎉

Verify Cluster

To make sure the Kubernetes cluster deployment was a success, log into the Control Plane node with the ubuntu username:

ssh ubuntu@control_plane_ip

Then check if all the nodes are connected to the K8s cluster:

kubectl get nodes

If the following prints, you're in the clear and ready to deploy workloads:

control1   Ready    control-plane,master   74m   v1.22.4
worker1    Ready    <none>                 72m   v1.22.4
worker2    Ready    <none>                 72m   v1.22.4

Contributing

If you like to add or change some of the code in this repo feel free to open a pull request and I will take look as soon as possible.

Tested On

Operating System	Virtual Machine	Bare Metal	CPU Architecture	Kubernetes Version
Ubuntu 20.04	Yes	No	x86	1.22
Ubuntu 22.04	Yes	No	x86	1.24

To-Do

• Test on Raspberry Pi
• Test on RHEL Distro
• Add Group Vars configuration file
• Test on Ubuntu Bare Metal

License

MIT