Mastering Kubernetes One Task at a Time— Creating A Kubernetes Cluster with Kubeadm using Terraform and AWS EC2

100% Automated. Ready, Fire, Cluster! A perfect playground to experiment and learn Kubernetes. Practice CKA and CKAD.

Table of Contents

1. Introduction
2. Clone Cluster GitHub Repo and Apply Terraform
3. Code Walk-through and Highlights
4. Conclusion

Introduction

Recently I had an opportunity to revisit my earlier work, This Code Works! — A Fully Functional Kubernetes Cluster with Kubeadm on AWS EC2.

Remarkably and predictably, following the AWS CLI code statements, I could spin up a Kubernetes cluster in less than 10 minutes, exactly as promised. That’s the good news. The bad news is that it is non-IaC and manual. As a result, it’s not a practical long-term solution.

As I am now actively preparing for the CKA and CKAD, there is an even urgent motivation for me to find a 100% IaC solution so that I can create and destroy a cluster with “a click of a button,” practice and experiment completely at my liberty with full control on all the nodes.

It took a little bit of work to get my bearing back on Terraform by referring back to my previous works, and by leveraging the prequel work combined with some inspiration from this timely blog, I am happy to announce that I now have a 100% Terraform-driven solution, and it works perfectly.

The code is in GitHub. So let’s clone and get crackin’!

Clone Cluster GitHub Repo and Apply Terraform

Below is the gist code showing from the beginning to the end of the Kubernetes infrastructure creation, necessary package deployment, and cluster configuration by running terraform on the GitHub code repo cloned in a local environment.

As expected, just firing terraform apply does the full job. After the terraform run, the final step is to ssh into the control plane node and confirm the cluster is up and running.

Code Walk-through and Highlights

Below are the normal terraform template files and some shell scripts under the kubeadm-scripts folder. Worth pointing out:

1. main.tf — Creates the EC2 nodes and does the necessary Security Group and Key Pair setup as necessary. The variable worker_node_count (default 2) in variables.tf file controls the number of worker nodes.
2. step-01-k8s-packages.sh — This script runs on all nodes via the remote-exec provisioner. See comments in the script to understand each step.
3. step-02-k8s-cp-init.sh — This script runs on the control plane node only, again via the remote-exec provisioner. The most important outcome of this script is the node join command.
4. step-03-k8s-join.sh — This is created during run time via the local-exec provisioned as part of the control plane instance creation. It contains the join command to be run on each worker node, obtained by running the kubeadm token create — print-join-command command on the control plane node. The script, once ready, runs on each worker node via the remote-exec provisioner.

terraform-k8s-kubeadm-ec2-cluster$ tree .
.
├── README.md
├── kubeadm-scripts
│   ├── step-01-k8s-packages.sh
│   ├── step-02-k8s-cp-init.sh
│   └── step-03-k8s-join.sh
├── main.tf
├── outputs.tf
└── variables.tf

Conclusion

Here I’ve presented the code and explained how to create a fully functional Kubernetes cluster with Kubeadm using Terraform and AWS EC2.

Because it is IaC using terraform, the first part of the post shows step-by-step commands to create the infrastructure and configure the Kubernetes cluster, all via terraform templates and run.

The second part highlights the essential components of the code and the mechanism by which the nodes are configured remotely from a local terraform-enabled environment.

Overall this 100% IaC automation of creating a Kubernetes cluster with Kubeadm is a highly efficient and improved solution compared to its predecessor, which did everything using AWS CLI.

The code should work as-is out of the box. If you encounter any issues, please let me know via comments. All the best!

If you benefited from reading the post, please 👏 a few times before parting, and help others by sharing it; I highly appreciate that!

Please follow to stay in touch, track, and be the first to get notified of all future writings on AWS Cloud, Containers, Kubernetes, and Machine Learning. Also, check all my stories on The AWS Way publication.