Deploy a kubernetes cluster with kubespray

Note

This guide is suitable for users with some experience using MyElits (Openstack) and Linux.

Kubernetes (k8s) is an open-source system for automating deployment, scaling, and management of containerized applications. You can use it to:

• Deploy your applications quickly and predictably.
• Scale your applications on the fly.
• Roll out new features seamlessly.
• Limit hardware usage to required resources only.

Kubespray is a tool that can be used to deploy Kubernetes on Openstack. It supports most popular Linux distributions and is highly customizable.

In this guide we will:

• Launch and configure a management server which will be used to run Kubespray.
• Deploy Kubernetes on three servers (one master and two nodes).

Management Server

Let's start by creating the management server in our Openstack project.

Name	kubespray
Source	Ubuntu 16.04 (Xenial Xerus)
Volume Size	10 GB
Flavour	1 CPU 1 GB
Networks	default-network
Security Groups	default, ssh
Key Pair	Your own keypair

 

Assign your server a floating IP so you can SSH to it.

SSH to the kubespray server and run:

$ ssh-keygen -t rsa
$ sudo apt-get update
$ sudo apt-get install python-pip
$ sudo pip install ansible kubespray shade

Copy your new SSH-key from ~/.ssh/id_rsa.pub and add it as a new key in MyELITS or copy your existing key to the kubespray server.

A configuration file will be created in ~/.kubespray.yml. Edit it to your liking, example:

# OpenStack options
os_auth_url: "https://lpi-api.elits.com:5000/v2.0"
os_username: "Your MyELITS email"
os_password: "Your MyELITS password"
os_project_name: "Your MyELITS project name"
os_region_name: "RegionOne" # RegionOne for Linkoping or Stockholm
masters_flavor: "general1-2" # 2 CPU 2 GB RAM
nodes_flavor: "general1-2" # 2 CPU 2 GB RAM
etcds_flavor: "general1-2" # 2 CPU 2 GB RAM
image: "Ubuntu 16.04 (Xenial Xerus)"
network: "default-network"
sshkey: "kubespray"
#

 After everything is configured, run:

# 1 master and 2 workers
$ kubespray openstack --masters 1 --nodes 2 --etcds 1

This will create three servers in MyELITS and configure kubespray in ~/.kubespray/.

Ubuntu 16.04 only ships with python3 while python2 is required on all hosts for ansible to work correctly. Therefore we use ansible to install it:

$ ansible -i ~/.kubespray/inventory/inventory.cfg all --sudo -m raw -a "apt-get update && apt-get install -y python python-simplejson"

Now it's time to deploy kubernetes on the nodes:

$ kubespray deploy

If everything was successful you should now see:
Kubernetes deployed successfuly

You can look in ~/.kubespray/inventory/inventory.cfg to see which of your servers is master.
SSH to the master and run:

$ kubectl get nodes
NAME                    STATUS    ROLES     AGE       VERSION
k8s-antagonism-161lhh   Ready     node      1h        v1.8.4+coreos.0
k8s-antagonism-b7jqrj   Ready     node      1h        v1.8.4+coreos.0
k8s-antagonism-df6eyf   Ready     master    1h        v1.8.4+coreos.0

If it's your first time running Kubernetes you can continue with this guide to deploy a stateless application.

Sources:

• https://kubernetes.io/
• https://kubernetes.io/docs/concepts/overview/what-is-kubernetes/
• https://github.com/kubernetes-incubator/kubespray