Kubernates the hardway using Vagrant

This is a vagrant project that follows this guide: Kubernates The Hard Way but using Vagrant instead of Google Cloud Engine.

This is not code to run on production environment, but it can help to understand the steps involved into configuring a Kubernates cluster.

Prerequisites

You need to have Vagrant installed in your machine. Also, once Vagrant is intalled you need to install Vagrant Triggers:

vagrant plugin install vagrant-triggers

Also, you need to create the following folders:

$ mkdir shared
$ mkdir ssh_keys

Inside the ssh_keys folder, generate a ssh key of type ed25519. Using Linux, Mac OS X or Cygwin in Windows:

$ mkdir ssh_keys
$ cd ssh_keys/
$ ssh-keygen -t ed25519 -f id_ed25519

The environment

By default this generates 5 VMs, you can change the number of worker and controller nodes just editing this two variables of the Vagrantfile:

total_mumber_of_controllers = 2
total_number_of_workers = 2

The schema is the following one. A client/load balance machine is going to be created with ip address: 10.0.0.200. Then several controllers, to a maximum of 9, controller-1 will have ip address 10.0.0.51, controller-2 will have ip address 10.0.0.52, and so on. Finally, a maximum of 49 worker nodes can be created, worker-1 will have ip address 10.0.0.101, worker-2 will have ip address 10.0.0.102 and so on.

You can ssh to any of this VMs using the hostname, for example:

vagrant ssh controller-3

Post Configuration

Once the cluster is running, we need to install our DNS pod, as indicated here: The DNS Cluster Add-on:

$ kubectl create -f kube-dns.yaml 
service "kube-dns" created
serviceaccount "kube-dns" created
configmap "kube-dns" created
deployment "kube-dns" created