Bootstrapping a highly available self managed K8s cluster in AWS

This repository demonstrates how to bootstrap a highly available self managed Kubernetes cluster in AWS using Kubeadm.

If you are looking for a production grade tool to provision self managed Kubernetes clusters, you can check out Claudie. It is open source and can provision multi-cloud and hybrid-cloud Kubernetes clusters (without using Kubernetes Cluster Federation).

How to run

Don't forget to create the ami/variables.auto.pkrvars.hcl and terraform.tfvars.hcl files.

First of all, create the custom AMI (Amazon Machine Image) using Hashicorp Packer, by executing this command :

cd ami && \
  packer init . && \
  packer build .

After the AMI is built and becomes active, Packer will output the AMI id. That'll be the value of args.ami_id in the Terraform module.

After provisioning everything, the Terraform module outputs :

• Public IP address of the Bastian Host
• Private IP addresses of the master nodes
• DNS name of the internal AWS ELB sitting in front of the master nodes

You will also find the kubeconfig.yaml and private-key.pem (contains SSH private key) files at /outputs.

To access the Kubernetes cluster, SSH into the Bastian Host using this command -

chmod 0400 ./outputs/private-key.pem
ssh -i ./outputs/private-key.pem ubuntu@bastian_host_public_ip

chmod 0400 private-key.pem
ssh -i private-key.pem ubuntu@first_master_node_private_ip

Then from the Bastian Host, SSH into the first master node. You can then access the Kubernetes cluster using kubectl.

References

• https://kubernetes.io/docs/setup/production-environment/tools/kubeadm/high-availability/
• AWS Docs