aws

Using Docker image

This is the easiest way to deploy an HA Kubernetes cluster on AWS with kubernautslabs/tk8 docker image.

Prerequisites

• Git

• Docker

• Existing SSH keypair in AWS

• AWS access and secret keys

git clone https://github.com/kubernauts/tk8
cd tk8
docker build -t kubernautslabs/tk8 . # optional, only if you'd like to build your own docker image
vi config.yaml --> provide your AWS access and secret key and an exsiting SSH keypair in AWS
docker run -it --name tk8 -v ~/.ssh/:/root/.ssh/ -v "$(pwd)":/tk8 kubernautslabs/tk8 sh
cd tk8
tk8 cluster init # clone kubernauts/kubespray
tk8 cluster aws --create # create the cluster infra using terraform
pip install -r kubespray/requirements.txt # not needed with kubernauts/kubespray
tk8 cluster aws --install
KUBECONFIG=./kubespray/inventory/awscluster/artifacts/admin.conf kubectl get pods --all-namespaces
tk8 cluster aws --destroy

Provision the cluster using CLI

Prerequisites

• Git

• Terraform

• Ansible

You can download the TK8 CLI for Linux and Mac OSX under releases here.

Make sure you have the sufficient permissions to create following resources in AWS

• VPC with Public/Private Subnets, and NAT Gateways in different Availability Zones

• EC2 instances used for bastion host, masters, etcd, and worker nodes

• AWS ELB in the Public Subnet for accessing the Kubernetes API from the internet

• IAM Roles which will be used with the nodes

Provide the AWS credentials in either of the following ways:

• AWS IAM Instance Profile

• Configuration and Credential Files

• Environment Variables. You will need to specify AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY.

• Config file

Clone the tk8 repo

git clone https://github.com/kubernauts/tk8
cd tk8

Adapt the config.yaml file to specify the cluster details. Example config:

aws:
   clustername: kubernauts
   os: centos # valid options are coreos/ubuntu/centos
   aws_access_key_id: # optional, see above
   aws_secret_access_key: # optional, see above
   aws_ssh_keypair: # needs to be an existing SSH keypair in AWS
   aws_default_region:
   aws_vpc_cidr_block : "10.250.192.0/18"
   aws_cidr_subnets_private : '["10.250.192.0/20","10.250.208.0/20"]'
   aws_cidr_subnets_public : '["10.250.224.0/20","10.250.240.0/20"]'
   aws_bastion_size : "t2.medium"
   aws_kube_master_num : 1
   aws_kube_master_size : "t2.medium"
   aws_etcd_num : 1
   aws_etcd_size : "t2.medium"
   aws_kube_worker_num : 2
   aws_kube_worker_size : "t2.medium"
   aws_elb_api_port : 6443
   k8s_secure_api_port : 6443
   kube_insecure_apiserver_address : "0.0.0.0"
   kubeadm_enabled: false # set to true to enable Kubeadm support

Once done run:

tk8 cluster init
tk8 cluster aws --create
tk8 cluster aws --install

Post installation the kubeconfig will be available at: ./kubespray/inventory/awscluster/artifacts/admin.conf

Do not delete the kubespray directory post installation as the cluster state will be saved in it.

Destroy the provisioned cluster

Make sure you are in same directory where you executed tk8 cluster init with the cloned kubespray directory.

To delete the provisioned cluster run:

tk8 cluster aws --destroy