Ansible Kubeadm Setup

Build a Kubernetes cluster using Ansible with kubeadm. The goal is easily install a Kubernetes cluster on machines running:

• Ubuntu 20.04
• Ubuntu 16.04
• CentOS 7
• RedHat 8

System requirements:

• Deployment environment must have Ansible
• Node with atleast 2GB RAM and 2vCPU

Usage

Add the system information gathered above into a file called hosts. You must have connectivity between all the nodes. We can use Multi-Cloud to launch the instance also either for master or slave node. For example:

[master]
#ip_of_master_node
65.2.35.218

[slave]
#ip_of_slave_nodes
65.1.148.10[2:5]

Edit remote_user and private_key_file in ansible.cfg before proceed to next steps.

Before continuing, edit group_vars/all.yml to your specified configuration.

For example, I choose to run crio instead of docker, and thus:

# container runtimes ('docker', 'crio')
container_runtime: crio

Requirements

Some modules in roles is also dependent on some other ansible galaxy collection. So run the requirements.yml first...

$ ansible-galaxy collection install -r requirements.yml
...
Process install dependency map
Starting collection install process
Installing 'community.general:2.5.1' to '/root/.ansible/collections/ansible_collections/community/general'

Playbook

- name: Install Necessary Software
  hosts: all
  vars_files:
    - group_vars/all.yml
  roles:
  -  role: runtime/{{ container_runtime }}-setup
  -  role: common
- name: Setup Master node
  hosts: master
  roles:
  -  role: master
  -  role: cni
- name: Setup Slave  node
  hosts: slave
  roles:
  -  role: node

Run this playbook for provision cluster

$ ansible-playbook role-kube.yml

Now, Verify cluster running using Kubectl command in master node:

$ kubectl get nodes