Skip to content

Latest commit

 

History

History

playbooks

README.md

Example Playbooks for Kubernetes cluster deployment with Fury

In this directory we provide example playbooks for deploying a Kubernetes cluster using Kubernetes Fury Distribution to on-premises virtual machines at version 1.23.12 and then how to upgrade it to 1.24.7. The process is the same for the other supported Kuberntes versions.

Requirements

To be able to run the examples, you need to have the following software installed:

  • ansible >= 2.8.0
  • furyagent
  • kubectl

One of the following supported Operating Systems on the target machines:

  • Ubuntu 20.04 LTS
  • RHEL 8
  • Rocky 8
  • Debian 12

Cluster Architecture

The cluster is composed of:

  • 2 HAproxy load balancers
  • 3 Master nodes
  • 3 Worker nodes

In the master nodes we also deploy etcd as a standalone systemd service.

Check the following files for a complete example:

  • hosts.yaml
  • haproxy.cfg

NOTE: all the cluster configuration is managed by hosts.ini and haproxy.cfg files.

Install phases

The install phases are ordered, and each playbook has the order in which it should be run in the file name.

Initialize PKI

First of all, we need to initialize CAs certificates for the etcd cluster and the Kubernetes components. The initialization of the PKI is done locally, using furyctl:

furyctl create pki

As a result, a ./pki directory with the following files will be created:

pki
├── etcd
│  ├── ca.crt
│  └── ca.key
└── master
   ├── ca.crt
   ├── ca.key
   ├── front-proxy-ca.crt
   ├── front-proxy-ca.key
   ├── sa.key
   └── sa.pub

Install the Container Runtime

In this step you can choose which container runtime you want to use:

  • containerd

Containerd

Run the 1.containerd.yml playbook with:

ansible-playbook 1.containerd.yml

Install the Load Balancer

To install the load balancer run the 2.load-balancer.yml playbook with:

ansible-playbook 2.load-balancer.yml

Provision Master and Worker Nodes

Now that all the prerequisites are installed, we can provision the Kubernetes master and worker nodes.

The 3.cluster.yml playbook needs some variables to be set in the hosts.ini file, double-check that everything is ok.

Run the playbook with:

ansible-playbook 3.cluster.yml

Upgrade cluster

In this folder there are two playbooks to upgrade the cluster to a new kubernetes version.

Change the hosts.ini with the version you want to upgrade to:

all:
  vars:
    kubernetes_version: '1.24.7'

NOTE: the kubernetes_version must be one of the versions available in the roles, i.e. supported by this installer.

IMPORTANT: all the nodes must be in Ready status before running the upgrade.

First you need to upgrade the control plane with the 55.upgrade-control-plane.yml playbook with:

ansible-playbook 55.upgrade-control-plane.yml

Then you need to upgrade the worker nodes with the 56.upgrade-worker-nodes.yml playbook with:

ansible-playbook 56.upgrade-worker-nodes.yml --limit worker1

Repeat this step foreach worker node in the cluster.

NOTE: you need to run the playbook with the --limit option to limit the nodes to upgrade. Why? Because the upgrade process will drain the node before upgrading it. Additionally, during the upgrade containerd will also be upgraded

Utilities

How to migrate from Docker to Containerd

To migrate from docker to containerd, there is an example playbook in this directory 99.migrate-docker-to-containerd.yml.

It must be executed one node at a time:

ansible-playbook 99.migrate-docker-to-containerd.yml --limit worker1