Project that facilitates the creation of a multinode Kubernetes cluster using Vagrant, Virtualbox and we running Ubuntu for study purposes.
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.gitignore
LICENSE
README.md
Vagrantfile
kubeadm-config.yaml
setup-cluster.sh

README.md

Kubernetes Vagrant

This project brings up 3 Virtualbox VMs running Ubuntu 16.04. Then we install kubeadm and it's dependencies on version 1.11.3 (1.12 are still breaking somethings). The Calico CNI will be set up for networking.

To use Weave CNI instead of Calico just invert the comments in setup-cluster.sh, where it mention calico we comment and where it mention weave just uncomment.

0. Install dependencies

Install Virtualbox, Vagrant, vagrant-cachier and vagrant-vbguest plugins for vagrant and kubernetes cli.

OSX with homebrew (https://brew.sh/)

brew cask install virtualbox
brew cask install vagrant
brew install kubernetes-cli

Windows with chocolatey

choco install virtualbox
choco install vagrant
choco install kubernetes-cli

Linux, FreeBSD, OpenBSD, others

You may know what to do. Links:

Vagrant plugins

vagrant plugin install vagrant-cachier vagrant-vbguest

1. Get the source

First we need to clone the project:

git clone https://github.com/wsilva/kubernetes-vagrant
cd kubernetes-vagrant

2. Set up the cluster

Just run the following shell script file or if you want just run each command and follow it

./setup-cluster.sh

3. Setting up kubconfig

You can merge the generated kubernetes-vagrant-config file with your $HOME/.kube/config file. And redo it everytime o set up the cluster again.

Or you can just export the following env var to point to both config files:

export KUBECONFIG=$HOME/.kube/config:kubernetes-vagrant-config

And then select the brand new vagrant kubernetes cluster created:

kubectl config use-context kubernetes-admin@k8s-vagrant

4. Shut down or reset

For shutting it down we just need to vagrant halt. When you need your cluster back just ./setup-cluster.sh again, it will be reprovisioned but way faster than the first time.