terraform-aws-eks
Deploy a full AWS EKS cluster with Terraform
What resources are created
- VPC
- Internet Gateway (IGW)
- Public and Private Subnets
- Security Groups, Route Tables and Route Table Associations
- IAM roles, instance profiles and policies
- An EKS Cluster
- EKS Managed Node group
- Autoscaling group and Launch Configuration
- Worker Nodes in a private Subnet
- bastion host for ssh access to the VPC
- The ConfigMap required to register Nodes with EKS
- KUBECONFIG file to authenticate kubectl using the
aws eks get-token
command. needs awscli version1.16.156 >
Configuration
You can configure you config with the following input variables:
Name | Description | Default |
---|---|---|
cluster-name |
The name of your EKS Cluster | eks-cluster |
aws-region |
The AWS Region to deploy EKS | us-east-1 |
availability-zones |
AWS Availability Zones | ["us-east-1a", "us-east-1b", "us-east-1c"] |
k8s-version |
The desired K8s version to launch | 1.13 |
node-instance-type |
Worker Node EC2 instance type | m4.large |
root-block-size |
Size of the root EBS block device | 20 |
desired-capacity |
Autoscaling Desired node capacity | 2 |
max-size |
Autoscaling Maximum node capacity | 5 |
min-size |
Autoscaling Minimum node capacity | 1 |
vpc-subnet-cidr |
Subnet CIDR | 10.0.0.0/16 |
private-subnet-cidr |
Private Subnet CIDR | ["10.0.0.0/19", "10.0.32.0/19", "10.0.64.0/19"] |
public-subnet-cidr |
Public Subnet CIDR | ["10.0.128.0/20", "10.0.144.0/20", "10.0.160.0/20"] |
db-subnet-cidr |
DB/Spare Subnet CIDR | ["10.0.192.0/21", "10.0.200.0/21", "10.0.208.0/21"] |
eks-cw-logging |
EKS Logging Components | ["api", "audit", "authenticator", "controllerManager", "scheduler"] |
ec2-key-public-key |
EC2 Key Pair for bastion and nodes | ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQD3F6tyPEFEzV0LX3X8BsXdMsQz1x2cEikKDEY0aIj41qgxMCP/iteneqXSIFZBp5vizPvaoIR3Um9xK7PGoW8giupGn+EPuxIA4cDM4vzOqOkiMPhz5XK0whEjkVzTo4+S0puvDZuwIsdiW9mxhJc7tgBNL0cYlWSYVkz4G/fslNfRPW5mYAM49f4fhtxPb5ok4Q2Lg9dPKVHO/Bgeu5woMc7RY0p1ej6D4CKFE6lymSDJpW0YHX/wqE9+cfEauh7xZcG0q9t2ta6F6fmX0agvpFyZo8aFbXeUBr7osSCJNgvavWbM/06niWrOvYX2xwWdhXmXSrbX8ZbabVohBK41 [email protected] |
You can create a file called terraform.tfvars or copy variables.tf into the project root, if you would like to over-ride the defaults.
How to use this example
NOTE on versions The versions of this module are compatible with the following Terraform releases. Please use the correct version for your use case:
version = 3.0.0 >
with terraform0.13.x >
version = 2.0.0
with terraform< 0.12.x
version = 1.0.4
with terraform< 0.11.x
Have a look at the examples for complete references You can use this module from the Terraform registry as a remote source:
module "eks" {
source = "WesleyCharlesBlake/eks/aws"
aws-region = "us-east-1"
availability-zones = ["us-east-1a", "us-east-1b", "us-east-1c"]
cluster-name = "my-cluster"
k8s-version = "1.17"
node-instance-type = "t3.medium"
root-block-size = "40"
desired-capacity = "3"
max-size = "5"
min-size = "1"
vpc-subnet-cidr = "10.0.0.0/16"
private-subnet-cidr = ["10.0.0.0/19", "10.0.32.0/19", "10.0.64.0/19"]
public-subnet-cidr = ["10.0.128.0/20", "10.0.144.0/20", "10.0.160.0/20"]
db-subnet-cidr = ["10.0.192.0/21", "10.0.200.0/21", "10.0.208.0/21"]
eks-cw-logging = ["api", "audit", "authenticator", "controllerManager", "scheduler"]
ec2-key-public-key = "ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQD3F6tyPEFEzV0LX3X8BsXdMsQz1x2cEikKDEY0aIj41qgxMCP/iteneqXSIFZBp5vizPvaoIR3Um9xK7PGoW8giupGn+EPuxIA4cDM4vzOqOkiMPhz5XK0whEjkVzTo4+S0puvDZuwIsdiW9mxhJc7tgBNL0cYlWSYVkz4G/fslNfRPW5mYAM49f4fhtxPb5ok4Q2Lg9dPKVHO/Bgeu5woMc7RY0p1ej6D4CKFE6lymSDJpW0YHX/wqE9+cfEauh7xZcG0q9t2ta6F6fmX0agvpFyZo8aFbXeUBr7osSCJNgvavWbM/06niWrOvYX2xwWdhXmXSrbX8ZbabVohBK41 [email protected]"
}
output "kubeconfig" {
value = module.eks.kubeconfig
}
output "config-map" {
value = module.eks.config-map-aws-auth
}
Or by using variables.tf or a tfvars file:
module "eks" {
source = "WesleyCharlesBlake/eks/aws"
aws-region = var.aws-region
availability-zones = var.availability-zones
cluster-name = var.cluster-name
k8s-version = var.k8s-version
node-instance-type = var.node-instance-type
root-block-size = var.root-block-size
desired-capacity = var.desired-capacity
max-size = var.max-size
min-size = var.min-size
vpc-subnet-cidr = var.vpc-subnet-cidr
private-subnet-cidr = var.private-subnet-cidr
public-subnet-cidr = var.public-subnet-cidr
db-subnet-cidr = var.db-subnet-cidr
eks-cw-logging = var.eks-cw-logging
ec2-key-public-key = var.ec2-key
}
IAM
The AWS credentials must be associated with a user having at least the following AWS managed IAM policies
- IAMFullAccess
- AutoScalingFullAccess
- AmazonEKSClusterPolicy
- AmazonEKSWorkerNodePolicy
- AmazonVPCFullAccess
- AmazonEKSServicePolicy
- AmazonEKS_CNI_Policy
- AmazonEC2FullAccess
In addition, you will need to create the following managed policies
EKS
{
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Action": [
"eks:*"
],
"Resource": "*"
}
]
}
Terraform
You need to run the following commands to create the resources with Terraform:
terraform init
terraform plan
terraform apply
TIP: you should save the plan state
terraform plan -out eks-state
or even better yet, setup remote storage for Terraform state. You can store state in an S3 backend, with locking via DynamoDB
Setup kubectl
Setup your KUBECONFIG
terraform output kubeconfig > ~/.kube/eks-cluster
export KUBECONFIG=~/.kube/eks-cluster
Authorize users to access the cluster
Initially, only the system that deployed the cluster will be able to access the cluster. To authorize other users for accessing the cluster, aws-auth
config needs to be modified by using the steps given below:
- Open the aws-auth file in the edit mode on the machine that has been used to deploy EKS cluster:
sudo kubectl edit -n kube-system configmap/aws-auth
- Add the following configuration in that file by changing the placeholders:
mapUsers: |
- userarn: arn:aws:iam::111122223333:user/<username>
username: <username>
groups:
- system:masters
So, the final configuration would look like this:
apiVersion: v1
data:
mapRoles: |
- rolearn: arn:aws:iam::555555555555:role/devel-worker-nodes-NodeInstanceRole-74RF4UBDUKL6
username: system:node:{{EC2PrivateDNSName}}
groups:
- system:bootstrappers
- system:nodes
mapUsers: |
- userarn: arn:aws:iam::111122223333:user/<username>
username: <username>
groups:
- system:masters
- Once the user map is added in the configuration we need to create cluster role binding for that user:
kubectl create clusterrolebinding ops-user-cluster-admin-binding-<username> --clusterrole=cluster-admin --user=<username>
Replace the placeholder with proper values
Cleaning up
You can destroy this cluster entirely by running:
terraform plan -destroy
terraform destroy --force