Brief

This Vagrant configuration let you deploy a basic, local Kubernetes cluster with the Weave-Net CNI plugin. Also it is possible to launch a highly available load balancer using HAProxy and KeepAlived, but it is disabled by default since it requires more horse power.

Note: Please don't hate me, but I tested it on Linux environment, and works perfectly. Windows can be more RAM and CPU greedy, so may be provisioning VMs never ends.

Deployment

Take a quick look on the variables at the top of the Vagrantfile. Those allows you to select how many master and worker nodes you want to deploy. Additionally, you can decide between deploy a HAProxy load balancer or not. It is disabled by default. Be familiarized with the values set and be sure that those matches with your capacity and intentions.

Keep in mind that more nodes or even enabling the load balancer will demand more resource power (Two VMs will be launched to run HAProxy and KeepAlived). It's recommended to plan and validate if you have enough resources in your machine before to deploy or modify defaults.

Once ready to deploy, hit:

vagrant up

You must use a mean to translate names to IPs (DNS) such as the hosts file or a DNS server. Take care of the IP used by the master-1 or the load balancer virtual IP, due you need to put it into your /etc/hosts file with your "cluster_endpoint_name", which is used on the generated kubeconfig file.

For instance, if you did enable the load balancer by use_loadbalancer = true, just use the value of the variable "keepalived_virtual_ip" and put it into your /etc/hosts file, translating to the value of "cluster_endpoint_name" variable, like this:

echo "192.168.56.100   my-kubernetes-cluster" | sudo tee -a /etc/hosts

If you use defaults, just hit:

echo "192.168.56.21   my-kubernetes-cluster" | sudo tee -a /etc/hosts

At the end of the deployment, Vagrant put the admin kubeconfig file on the working directory to connect to the new Kubernetes cluster. You can choose one of the following options to connect locally to the deployed cluster:

Option A

Set the configuration using the KUBECONFIG envvar and kubeconfig the file into the current directory. Make sure you are still inside of this repo when launch the command:

export KUBECONFIG=$(pwd)/kubeconfig

Note:

This leverages the Vagrant capability of mount the Vagrantfile parent folder inside of the VM on /vagrant path. On the Vagrantfile, I copy the generated kubeconfig to the /vagrant/ folder, which makes it available for you into the repo directory.

Option B

Move the kubeconfig file to the $HOME/.kube folder and set the KUBECONFIG envvar to use it. This lets you to have the kubeconfig file into a recommended path without impact your current $HOME/.kube/config file, that may be present.

[[ ! -d ~/.kube ]] && mkdir ~/kube
cp ./kubeconfig ~/.kube/vagrant-cluster.config
export KUBECONFIG=~/.kube/vagrant-cluster.config

Get the nodes status once you configured properly your environment:

kubectl get nodes

Visit the Kubernetes Cheat Sheet for tips!