Highly Available Kubernetes Cluster Demo

This repository contains all of the code needed to stand up a highly available Kubernetes cluster in AWS.

The cluster has the following requirements/specifications:

• 3 Master Nodes
• 3 Worker Nodes
• Uses all 3 eu-west-1 (Ireland) availability zones (eu-west-1a, eu-west-1b & eu-west-1c)
• Uses the Route53 DNS zone aws.lb-dev.io
• Uses t3.medium instances for masters and workers
• Not accessible from the internet
• Uses a bastion server
• Access is locked down to the specified IP
• Uses kube2iam for assuming IAM roles
• Uses external-dns for automatically creating Route53 records
• Uses AWS ALB for the Ingress controller
• Weave network traffic is encrypted
• EBS volumes for persistent storage are encrypted

Requirements

To use this repository you will need the following installed:

• kubectl - Kubectl
• kops - kops
• jq - jq
• openssl - This is included with most Linux distributions
• Helm - Helm
• AWS Account with awscli configured

These will need to be installed on your machine before running the scripts.

Preflight Checks

You will need to make changes to the following files in order to successfully spin up the cluster:

• ./create-cluster.sh - You will need to change the IP for --admin-access and apply any tags you wish. You can also change the AWS region and availability zones in here.
• ./bootstrap.sh - You will need to change the values for NAME and KOPS_STATE_STORE and also change the values for the Route53 setup and the S3 bucket setup. Make sure the name in the env var for the s3 bucket matches the bucket you create.

NOTE:- You will also need to look for any mention of aws.lb-dev.io and replace this with the domain you use in Route53.

Helm Certificate Setup

In order for the Helm/Tiller components to work correctly you will need to generate the required certificates.

The commands for this are located in cluster-bootstrap/scripts/helm-ssl. You will need to run these to generate the keys and certificates for Helm/Tiller to correctly work.

Usage

To start the cluster run the scripts in the following order:-

• source ./bootstrap.sh
• ./create-cluster.sh
• ./post-setup.sh

NOTE:- You need to run bootstrap.sh using source so that the environment variables are set correctly for kops to function.

Post Setup

Once the post-setup.sh script has completed you will then have a fully functioning HA Kubernetes cluster.

Included is a sample app that you can deploy to check the ingress and external DNS are functioning correctly, just run the following commands:

• kubectl create -f sample-app.yaml
• kubectl create -f sample-ingress.yaml