Multi-Node Cluster Setup with Kind

Learning and testing multi-node behavior needs a cluster, but real multi-machine clusters are expensive and slow to provision. We need a cheap, fast, throwaway multi-node cluster that runs entirely on a laptop.

Solution: Run a real multi-node cluster locally with kind (each node is a Docker container) to learn and test without cloud cost.

Instead of leasing real waterfront and pouring concrete, you build a working model port on your desk: tiny docks and a control office made of toy blocks that genuinely move cargo. kind does exactly that, standing up a real multi-node cluster on your laptop where every node is just a Docker container. The model harbor master's office is the control-plane node, the toy model docks are the worker nodes, and the desk and power strip holding it all up are the Docker engine that hosts every piece.

The same cluster entities appear in every day's notes; the diagram below shows where this day's topic fits.

Graph legend — each node maps to a piece of the kind cka cluster:

Kind = Kubernetes IN Docker. It runs each Kubernetes "node" as a Docker container. Perfect for local learning/CI — a full multi-node cluster on a laptop.

Graph legend — each node maps to a container created by the cka kind config:

You need three tools: kind to build the cluster, kubectl to talk to it, and a running Docker engine to host the node containers.

# kind
brew install kind          # or: go install / curl binary
# kubectl
brew install kubectl
docker --version           # Docker must be running

A small YAML file declares how many nodes you want and their roles (one control-plane, two workers here).

kind-config.yaml:

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

One command reads the config and boots the cluster; kind then wires up your kubectl context automatically.

kind create cluster --name cka --config kind-config.yaml

# kind auto-configures kubectl context to this cluster
kubectl cluster-info --context kind-cka
kubectl get nodes

Expected:

NAME                 STATUS   ROLES           AGE   VERSION
cka-control-plane    Ready    control-plane   1m    v1.xx
cka-worker           Ready    <none>          1m    v1.xx
cka-worker2          Ready    <none>          1m    v1.xx

A context binds kubectl to a specific cluster, user, and namespace — switch contexts to point commands at a different cluster.

kubectl config get-contexts
kubectl config current-context
kubectl config use-context kind-cka

Graph legend — each node maps to the kind-cka context wiring:

Everyday housekeeping commands for listing, deleting, and peeking at the node containers.

kind get clusters            # list clusters
kind delete cluster --name cka
docker ps                    # see the node containers

Pass a specific node image to lock the cluster to the Kubernetes version you want to practise against.

kind create cluster --name cka \
  --image kindest/node:v1.30.0 --config kind-config.yaml

One config file to a ready multi-node cluster: kind boots node containers, bootstraps Kubernetes, and wires up your context.

Graph legend — each node maps to a step from config to a ready cka cluster:

• Kind nodes are Docker containers, not VMs — fast and disposable.
• One config file defines control-plane + worker roles.
• kubectl talks to the cluster via the context in your kubeconfig.

• [ ] Installed kind + kubectl, Docker running
• [ ] Created a 3-node cluster from config
• [ ] kubectl get nodes shows control-plane + 2 workers Ready
• [ ] Switched contexts and deleted a cluster