Ansible playbook for kubeadm-based Kubernetes cluster installation on Linux

---

This Ansible playbook automates the installation and configuration of a Kubernetes cluster on Linux, supporting both single control plane and HA control plane setups, using the latest stable Kubernetes release.

While Kubespray provides extensive features and customization options, this playbook remains lightweight and simple, making it an ideal choice for quickly setting up a development or testing Kubernetes environment on Linux.

Suitable Environment: Development & Testing

System Requirements: Minimum 2 GB RAM & 2 vCPU

Supported Platforms: Baremetal, Virtual Machines, Cloud Instances

Supported Linux distributions:

• RedHat-based ( Fedora, RHEL, Rocky Linux, Almalinux, Oracle Linux )
• Debian-based ( Debian, Ubuntu )
• SUSE-based ( OpenSUSE, SLES )

Prerequisites:

• Please install Ansible on the machine where you plan to run the playbook if you haven’t done so already.
• Prepare the cluster nodes by installing any of the above mentioned supported Linux distributions, even with a minimal installation.
• Please ensure that you have DNS set up that resolves all the involved hosts, or update the host files on all hosts with the necessary entries for each involved host.
• Create a common Linux user on all cluster nodes, which will be used for the cluster installation.
• Enable passwordless SSH authentication from the Ansible host to all cluster nodes using the common user created earlier.
• Ensure the common user has passwordless sudo privileges on all cluster nodes.
• For HA cluster setups, ensure that the control plane endpoint is configured via a load balancer such as NGINX, HAProxy, or any load balancer of your choice.

The main playbook installs and configures the latest stable versions of the following required components.

• Container orchestrator: kubernetes
• Container runtime: containerd
• Low-level container runtime: runc ( dependency for containerd )
• CNI plugin: calico

Optional playbooks are available to install the following components once the cluster is ready.

• csi-driver-nfs
• csi-driver-smb
• MetalLB LoadBalancer

---

Step-by-Step Workflow:

Step 1) Copy and execute the below command snippet to extract the tarball for the most recent stable release of this Ansible project.

curl -sSL https://github.com/Muthukumar-Subramaniam/install-k8s-on-linux/releases/latest/download/inst-k8s-ansible.tar.gz | tar -xzvf - && cd inst-k8s-ansible

Step 2) Update the host-control-plane file with the necessary hostnames.

Use a single control plane node for a single control plane setup. For HA cluster setups, ensure a minimum of 3 control plane nodes, and always use an odd number of nodes.

Single Control Plane Setup

HA Control Plane Setup

Additional Step for HA Control Plane Setup

• Update the file** control-plane-endpoint with the endpoint behind the load balancer that has all the control planes in the backend pool.

• Port configuration:
  If only <FQDN of control-plane-endpoint> is provided, the default port 6443 will be used.
  Alternatively, provide a specific port as <FQDN of control-plane-endpoint>:<port-number>.

  

  ( Or )

  

Step 3) Update the host-workers file with the necessary hostnames.

Step 4) Update the pod-network-cidr file with the desired pod network CIDR.

• Only private IP addresses, as defined in RFC 1918 are allowed.
• The deployment is configured to accept CIDR prefixes exclusively within the /16 to /28 range.
• Ensure that the selected CIDR prefix does not conflict with any existing networks in your infrastructure.
• Choose a CIDR prefix that provides sufficient address space for your cluster.

Step 5) Run the setup.py script to prepare the environment for the Ansible playbook.

./setup.py

Step 6) Run the playbook if the setup.py script completes successfully.

./inst-k8s-ansible.yaml

Expected Outcome:

Great work! Your cluster is now ready to use.

Optional Step 1) To install CSI NFS Driver for the kubernetes cluster if required.

./optional-k8s-csi-nfs-driver.yaml

Expected Outcome:

Optional Step 2) To install CSI SMB Driver for the kubernetes cluster if required.

./optional-k8s-csi-smb-driver.yaml

Expected Outcome:

Optional Step 3) To install MetalLB loadbalancer for the kubernetes cluster if required.

Note: Please make sure to change the address pool range in the playbook as per your environment and requirement.

./optional-install-metallb.yaml

Expected Outcome:

That's all for now! Your trust and engagement means a lot, and we hope you find the playbook useful.

Kindly note:

• This playbook is a useful resource for experimenting with Kubernetes and can be customized to meet your specific requirements.
• The playbook utilizes the GitHub API to fetch the current stable versions of all required software components.
• Compatible with a wide range of Linux distributions.
• Your feedback and contributions are invaluable to the success of this project.
• Please report any bugs, suggest new features, or contribute directly to the codebase.

Have lots of fun!