NOTE: This content is just for testing and experimentation. It can cause many security issues, so do not use it in a production environment.

Some days ago, a question came up in a Slack channel at Clastix, the company I currently work for: Could I launch a kind cluster on a remote machine? My immediate thought was yes, you could, but let’s see how.

Kind is a magnificent tool for people who develop for Kubernetes or test their software directly on Kubernetes. Every Kind node is a Docker container, and this is the key to being able to launch remote Kind clusters.

First, we need a remote machine. If you already have one, you can skip the next section and go directly to the Docker section.

Virtual Machine

For this experiment, a virtual machine will be used as the remote machine. You can use whichever system you are most comfortable with, but the example is based on Multipass to use KVM on a Linux machine.

Launch Virtual Machine

ubuntu@local:~$ multipass launch --name remote

List Virtual Machines

ubuntu@local:~$ multipass list                   
Name                    State             IPv4             Image
remote                  Running           10.192.157.65    Ubuntu 20.04 LTS

Log into the Virtual Machine

ubuntu@local:~$ multipass shell remote
Welcome to Ubuntu 20.04.3 LTS (GNU/Linux 5.4.0-91-generic x86_64)
ubuntu@remote:~$

Now we have a remote Linux machine running Ubuntu 20.04.

Docker

The Docker daemon is a server application that exposes an API to work with. By default, the Docker daemon is only exposed through a UNIX socket, but we can also expose a TCP socket by modifying the configuration.

First, we need Docker.

Installation

ubuntu@remote:~$ sudo apt update
ubuntu@remote:~$ sudo apt install -y docker.io

Configuration

The easiest way to manage Docker’s configuration is through the file /etc/docker/daemon.json, where we are going to expose the daemon on the remote machine’s IP. We will also keep the UNIX socket so we can still use the docker CLI from the remote machine:

{
  "debug": true,
  "hosts": ["tcp://10.192.157.65", "unix:///var/run/docker.sock"]
}

Then check that the ExecStart field in the file /lib/systemd/system/docker.service looks like this:

ExecStart=/usr/bin/dockerd  --containerd=/run/containerd/containerd.sock

Finally, restart the service and add your user to the docker group to avoid using sudo (this takes effect after you log out and back in):

ubuntu@remote:~$ sudo systemctl daemon-reload
ubuntu@remote:~$ sudo systemctl restart docker
ubuntu@remote:~$ sudo gpasswd -a $USER docker

Now we can work with Docker:

ubuntu@remote:~$ docker pull kindest/node:v1.23.0
v1.23.0: Pulling from kindest/node
cef70b0687d2: Pull complete 
1a0fe9cf114d: Pull complete 
Digest: sha256:49824ab1727c04e56a21a5d8372a402fcd32ea51ac96a2706a12af38934f81ac
Status: Downloaded newer image for kindest/node:v1.23.0
docker.io/kindest/node:v1.23.0

ubuntu@remote:~$ docker images
REPOSITORY     TAG       IMAGE ID       CREATED      SIZE
kindest/node   v1.23.0   b3dd68fe0a8c   2 days ago   1.46GB

Let’s go back to our local machine.

Pointing to the Remote Docker Daemon

The Docker CLI uses the environment variable DOCKER_HOST to build its requests, so:

export DOCKER_HOST="tcp://10.192.157.65"

Now all the docker cli requests will go to the remote machine:

ubuntu@local:~$ docker images
REPOSITORY     TAG       IMAGE ID       CREATED      SIZE
kindest/node   v1.23.0   b3dd68fe0a8c   2 days ago   1.46GB

Kind

The first step is to install Kind. Check how to install it here.

Check Kind Version

ubuntu@local:~$ kind --version
kind v0.11.1

Launch a Remote Kind Cluster

Kind offers a very flexible configuration, but in our case we will focus on the API Server.

ubuntu@local:~$ cat <<EOF | kind create cluster --image=kindest/node:v1.23.0 --name remote --config=-
kind: Cluster
apiVersion: kind.x-k8s.io/v1alpha4
networking:
  apiServerAddress: "10.192.157.65"
  apiServerPort: 6443
EOF
Creating cluster "remote" ...
 ✓ Ensuring node image (kindest/node:v1.23.0) 🖼
 ✓ Preparing nodes 📦  
 ✓ Writing configuration 📜 
 ✓ Starting control-plane 🕹️ 
 ✓ Installing CNI 🔌 
 ✓ Installing StorageClass 💾 
Set kubectl context to "kind-remote"
You can now use your cluster with:

kubectl cluster-info --context kind-remote

Have a question, bug, or feature request? Let us know! https://kind.sigs.k8s.io/#community 🙂

Check your current context

ubuntu@local:~$ kubectl config current-context
kind-remote

Check the Nodes

As we said, Kind’s nodes are Docker containers, so let’s check them:

ubuntu@local:~$ docker ps
CONTAINER ID        IMAGE                  COMMAND                  CREATED             STATUS              PORTS                          NAMES
e9e6723bf875        kindest/node:v1.23.0   "/usr/local/bin/entr…"   2 minutes ago       Up 2 minutes        10.192.157.65:6443->6443/tcp   remote-control-plane
ubuntu@local:~$ kubectl get nodes
NAME                   STATUS   ROLES                  AGE     VERSION
remote-control-plane   Ready    control-plane,master   2m36s   v1.23.0

Deploy Applications

Kind is ready to run your applications.

ubuntu@local:~$ kubectl run nginx --image=nginx:alpine
pod/nginx created

ubuntu@local:~$ kubectl get po
NAME    READY   STATUS    RESTARTS   AGE
nginx   1/1     Running   0          22s

Let’s check the app:

ubuntu@local:~$ kubectl port-forward nginx 8080:80 &
[1] 587897
Forwarding from 127.0.0.1:8080 -> 80
Forwarding from [::1]:8080 -> 80

ubuntu@local:~$ curl localhost:8080
Handling connection for 8080
<!DOCTYPE html>
<html>
<head>
<title>Welcome to nginx!</title>
<style>
html { color-scheme: light dark; }
body { width: 35em; margin: 0 auto;
font-family: Tahoma, Verdana, Arial, sans-serif; }
</style>
</head>
<body>
<h1>Welcome to nginx!</h1>
<p>If you see this page, the nginx web server is successfully installed and
working. Further configuration is required.</p>

<p>For online documentation and support please refer to
<a href="http://nginx.org/">nginx.org</a>.<br/>
Commercial support is available at
<a href="http://nginx.com/">nginx.com</a>.</p>

<p><em>Thank you for using nginx.</em></p>
</body>
</html>