Kubernetes in Docker

Quiz

As Kubernetes becomes the backbone of modern infrastructure, many of us look for efficient ways to spin up local clusters for development, experimentation, and testing. Minikube is one popular choice, but there's another powerful, lightweight alternative worth knowing about: KinD (Kubernetes in Docker).

KinD lets us run Kubernetes clusters in Docker containers, making it incredibly fast and resource-efficient. In this chapter, weâll walk through the steps to install KinD, spin up a cluster, interact with it, and deploy a basic application â all on our local machine.

Why KinD?

Letâs take a quick look at why KinD is important:

Designed for local development and CI pipelines.
Lightweight and fast â runs on Docker containers.
Supports multi-node clusters.
Actively maintained by the Kubernetes SIG community.
No need for VirtualBox or VMs â just Docker.

Prerequisites

Before we get started, letâs make sure we have everything we need:

A modern Linux, macOS, or Windows machine with at least:

4 GB of RAM
2+ CPU cores
Internet connectivity (to download KIND images)

Docker installed and running.
Kubernetes Cluster installed and running.

Install KinD

The easiest way is to grab the binary using curl:

$ curl -Lo ./kind https://kind.sigs.k8s.io/dl/v0.22.0/kind-linux-amd64
$ chmod +x ./kind
$ sudo mv ./kind /usr/local/bin/kind

Output

100    97  100    97    0     0    814      0 --:--:-- --:--:-- --:--:--   815
100 6245k  100 6245k    0     0  5972k      0  0:00:01  0:00:01 --:--:--  121M

To confirm the installation:

$ kind version

Output

kind v0.22.0 go1.20.13 linux/amd64

Create a Cluster

With Docker running in the background, we can now create our first KinD cluster. Simply run:

$ kind create cluster

Output

Creating cluster "kind" ...
â Ensuring node image (kindest/node:v1.29.2) ð¼ 
â Preparing nodes ð¦  
â Writing configuration ð 
â Starting control-plane ð¹ï¸ 
â Installing CNI ð 
â Installing StorageClass ð¾ 
Set kubectl context to "kind-kind"

To see if the cluster is running:

$ kubectl get nodes

Output

NAME                 STATUS   ROLES                  AGE   VERSION
kind-control-plane   Ready    control-plane          1m    v1.29.2

Create a Multi-Node Cluster (Optional)

If we want to simulate a real-world cluster with multiple nodes, KinD lets us define it using a config file.

Create a file called kind-multinode.yaml with the following lines:

kind: Cluster
apiVersion: kind.x-k8s.io/v1alpha4
nodes:
- role: control-plane
- role: worker
- role: worker

Then create the cluster:

$ kind create cluster --name my-multinode-cluster --config kind-multinode.yaml

Output

Creating cluster "my-multinode-cluster" ...
â Ensuring node image (kindest/node:v1.29.2) ð¼ 
â Preparing nodes ð¦  
â Writing configuration ð 
â Starting control-plane ð¹ï¸ 
â Installing CNI ð 
â Installing StorageClass ð¾ 
Set kubectl context to "kind-my-multinode-cluster"

Check the nodes

To check the status of your cluster nodes, run:

$ kubectl get nodes

Output

NAME                             STATUS   ROLES                  AGE   VERSION
kind-my-cluster-control-plane    Ready    control-plane          2m    v1.29.2
kind-my-cluster-worker-1         Ready    worker                 2m    v1.29.2
kind-my-cluster-worker-2         Ready    worker                 2m    v1.29.2

Deploy an Application

Letâs deploy a simple NGINX app to test our setup:

$ kubectl create deployment nginx --image=nginx

Output

deployment.apps/nginx created

Expose the deployment:

$ kubectl expose deployment nginx --port=80 --type=NodePort

Output

service/nginx exposed

Get the Service

$ kubectl get svc

Output

NAME         TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)          AGE
kube-dns     ClusterIP   10.96.0.10      <none>        53/UDP,53/TCP    2m
nginx        NodePort    10.96.0.175     <none>        80:32123/TCP     1m

Access the Service

To access the service, we need the Docker container's IP:

$ docker exec -it kind-control-plane curl localhost:32123

Output

<!DOCTYPE html>
<html>
<head>
   <title>Welcome to nginx!</title>
</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>

Managing KinD Clusters

List All Clusters:

$ kind get clusters

Output

kind
my-multinode-cluster

Delete a Cluster:

$ kind delete cluster

Output

deleting cluster "kind"

Or delete a specific one:

$ kind delete cluster --name my-multinode-cluster

Output

deleting cluster "my-multinode-cluster"

Load Docker Images into KinD

If we build an image locally and want to use it inside KinD:

$ kind load docker-image my-custom-image:latest

Output

Loaded image: my-custom-image:latest

Useful Tips

KinD is great for CI/CD: You can spin up and tear down clusters quickly during test pipelines.
Great for testing Kubernetes releases, as you can specify the version when creating a cluster:

$ kind create cluster --image kindest/node:v1.29.0

KinD is not meant for production â itâs a dev/test tool. You can also install Ingress controllers, MetalLB, and other components manually in KinD clusters.

Troubleshooting

Docker Not Running? KinD wonât work unless Docker is up and running. Make sure to start the Docker daemon.

kubectl context mismatch? KinD automatically updates the kubeconfig file. To switch contexts:

$ kubectl config use-context kind-kind

Conclusion

KinD is a fantastic tool when we want quick access to a Kubernetes environment on our local machine. Itâs lightweight, easy to configure, and fits perfectly into our dev and CI workflows. Using Docker containers, we avoid the overhead of VMs, making cluster startup and teardown incredibly fast.

Whether weâre building controllers, testing applications, or working on Kubernetes itself, KinD gives us a production-like environment locally â no VMs or cloud needed.

Print Page