K8s: Vagrant, Ansible Setup

Overview

This repo is an easy way to get a 3-Node Kubernetes cluster stood up using Vagrant, Ansible and Kubeadm.

A simple vagrant up once pre-requisites are installed, will get you all the way to a fully running cluster and kubectl ready to go.

Pre-requisites

Please get the following installed first:

• VirtualBox

# add /etc/apt.sources.list
deb [arch=amd64] https://download.virtualbox.org/virtualbox/debian <mydist> contrib

# add keys && verify (after)
wget -q https://www.virtualbox.org/download/oracle_vbox_2016.asc -O- | sudo apt-key add -
wget -q https://www.virtualbox.org/download/oracle_vbox.asc -O- | sudo apt-key add -

# update && install
sudo apt-get update
sudo apt-get install virtualbox-6.1

• Vagrant
• Ansible

install libarchive-tools to get tool bsdtar, which is required but not supported in Ubuntu 19+: sudo apt install -y libarchive-tools.

Deploy

In the rep folder, run: vagrant up

This will do the following:

1. create the k8s-control ubuntu VM Node.
2. install docker, kubernetes, kubeadm and init the cluster in k8s-control
3. create each worker-N ubuntu VM Node.
4. install docker, kubernetes, kubeadm and join each worker to k8s-control Node.

Check

Once it's all up, check for errors, then check with commands: vagrant status and you should see:

Current machine states:

k8s-control                running (virtualbox)
worker-1                    running (virtualbox)
worker-2                    running (virtualbox)

This environment represents multiple VMs. The VMs are all listed
above with their current state. For more information about a specific
VM, run `vagrant status NAME`.

Jump into a node with: vagrant ssh k8s-control or vagrant ssh node-1

Check your cluster:

vagrant@k8s-control:~$ kubectl get nodes
NAME          STATUS   ROLES                  AGE   VERSION
k8s-control   Ready    control-plane,master   87m   v1.23.5
worker-1      Ready    <none>                 83m   v1.23.5
worker-2      Ready    <none>                 80m   v1.23.5

KUBECONFIG

Access your cluster via kubectl from the k8s-control Node.

(TODO) - access the cluster from your host computer using the ~/.kube/config file from k8s-control.

troubleshooting

if you see this error when you run: journalctl -xeu kubelet

64] "Docker Info" dockerInfo=&{ID:JECY:YHJK:J7PW:2LCZ:B64F:M6PQ:BBXC:ORXL:HNRW:UXKT:X4CO:LM4D Containers:0 ContainersRunning:0 ContainersPaused:0 ContainersStopped:0 Images:7 Driver:ov
led to run kubelet" err="failed to run Kubelet: misconfiguration: kubelet cgroup driver: \"systemd\" is different from docker cgroup driver: \"cgroupfs\""
tatus=1/FAILURE

it means you need /etc/docker/daemon.json with this config in it:

{
  "exec-opts": ["native.cgroupdriver=systemd"]
}

then restart docker: systemctl restart docker

References

• k8s blog
• k8s kubeadm docs
• docker docs
• ansible blockinfile
• vagrant ubuntu/xenial64