Terraform Kube Cluster EC2

Terraform code to create a kube cluster using ec2. To be used with kubeadm or kubespray

Prerequisites

• Terraform >= 1.0 # This can be updated in provider.tf using required_version key

Resources Setup

• 1 VPC (Uses AWS VPC module)

  • 3 private subnets and 3 public subnets (Configurable)
  • Single NAT GW

• 1 Master Node

  • t3.medium

• 1 Worker node - Can be configured using worker-count key in variables.tf

  • t2.micro

• 3 Security groups

  • Configured using kubernetes documentation

  • Only the user/system from which the terraform apply is run, can connect to the instances created.

NOTE

The worker nodes do not have public IPs default (Can be changed).
In order to connect to them, use ssh forwarding concept

Usage

• Update the key_name key under variables.tf or pass the value at run time, to make use of an already existing ssh to connect to the instances.

• To initialise

      terraform init

• To run a plan

      terraform plan

• To deploy infrastructure

      terraform apply

• To destroy infrastructure

      terraform destroy

If automating, pass --auto-approve flag to apply and destroy commands

Optional

• By default, the backend for storing terraform state is the project directory/local system
• If you want to use S3 as your backend store, add the following to the terraform block in provider.tf

    backend "s3" {
        bucket  = <bucket-name>
        key     = <path/to/tf.state>
        region  = <region>
    }