terraform-aws-eks-cluster

Preface

This project is designed to be an example of standing up an EKS Cluster from scratch. It is not intended for production use, but with some minor modifications can be configured to be production ready. Removing the VPC definition(or keeping it!), and reconfiguring the Spot Fleet to use larger instances as well as autoscaling would make it more akin to a production cluster.

Pre-requisites

• Terraform Version 0.11.8 installed
• AWS Account
• kubectl installed
  • https://docs.aws.amazon.com/eks/latest/userguide/install-kubectl.html
• aws-iam-authenticator plugin for kubectl installed
  • https://github.com/kubernetes-sigs/aws-iam-authenticator
• Terraform user created in account with Admin permissions
  • Ensure that you have your API tokens set correctly in $HOME/.aws/credentials
• S3 Bucket created for storing state
• Dynamodb table created for locking state
  • If you do not need to lock your terraform run you can remove this from the backend

Configuration

Ensure that you modify the backend block in terraform to work with your environment. This is the first terraform block located at

./config.tf

This project is a kitchen sink project. It have every component required for standing up an EKS Cluster, it includes:

• Creating a VPC with:
  • 3 private subnets
  • 3 public subnets
  • NAT Gateway
  • Internet Gateway
• Creates an EKS Master Cluster with all required IAM Permissions
• Creates a Spot Fleet with Userdata configured to automatically join nodes to EKS

All that is necessary is to modify your the cluster name and a the allowed inbound IP address variables. The IP variable is intended to whitelist your IP address for connecting to the new cluster. Other modifications can be made to subnet IP allocations and names as necessary.

Once the necessary variables have been updated, run the following command:

$ terraform init
$ terraform apply

After this completes you will have two outputs, config-map-aws-auth and kubconfig. Place the kubeconfig in your $HOME/.kube/config. Verify that you can connect to the newly created cluster by running:

$ kubectl cluster-info

You should see a message similar to the following:

Kubernetes master is running at https://BDBC3C10C30CE483C169F0D7B932DEBC.yl4.us-east-1.eks.amazonaws.com

After you've verified that the cluster is running save the output labeled config-map-aws-auth to config-map-aws-auth.yaml. Then run the following command against it:

$ kubectl -f apply config-map-aws-auth.yaml

Finally run:

$ kubectl get-nodes --watch

And you should see the worker nodes joining the cluster.

#Cleaning Up The EKS master node is pretty expensive so I recommend not running it all the time unless you have a lot of cash to burn. When you're finished experimenting simply run:

$ terraform destroy

This will delete and cleanup all the resources created by this project.