YashIndane/multinode-k8s-cluster-on-AWS

Ansible roles to configure 🕸️ K8S multinode cluster on AWS

Kubernetes multinode cluster over AWS

Link to article -> Article

Ansible Roles

Total 3 roles have been used to achieve this, k8s_master_conf, k8s_worker_conf & launch_instances.

(Note- You don't have to create any inventory file, it get created automatically)

Prerequisites

Make the private key (.pem), protected by -

$ sudo chmod 400 <PRIVATE-KEY.pem>

Copy the ansible.cfg in this repository to /etc/ansible, and configure it accordingly.

$ sudo mkdir /etc/ansible && sudo mv ansible.cfg /etc/ansible

Launch the EC2 nodes -

$ ansible-playbook setup.yml

Launching the EC2-instances

Description for role variables used in launch_instances role ->

Variable	description
reg	region
img_id	AMI id
i_type	EC2 instance type
vpcsi	vpc-subnet id
key	private key name for instance EC2
sg_group	security group name
kube_master_name	name of EC2 instance for master node
aws_ak	AWS access key
aws_sk	AWS secret key
vol_size	volume size for EC2 instance (in Gib)
kube_worker_name1	name of EC2 instance for worker node-1
kube_worker_name2	name of EC2 instance for worker node-2

Write correct device name according to image used in launch_instances role.

(Note- by default it is configured to use device name as /dev/xvda)

Have minimum 8 Gib of volume for each node.

inside k8s_master_config role mention your pod network CIDR.

(Note- by default it is configured to use flannel and pod network cidr as 10.244.0.0/16)

Configuring Kubernetes Cluster on launched EC2 nodes

Configuring master node -

$ ansible-playbook setup-master.yml

Configuring worker nodes -

$ ansible-playbook setup-worker.yml

Joining worker nodes to master

Login to master node and get the join command -

$ sudo kubeadm token create --print-join-command

Now login to the worker nodes and execute the joining command on each of them.