Raspberry Pi Kubernetes K3s: Lightweight Cluster Setup

Kubernetes is the industry standard for container orchestration, and K3s makes it possible to run a fully functional cluster on Raspberry Pi hardware. Whether you are learning for career growth or building a production-grade home lab, a Pi-based K3s cluster is the most cost-effective way to master Kubernetes in India.

What is K3s and Why Use It on Raspberry Pi

K3s is a lightweight, certified Kubernetes distribution designed for resource-constrained environments. Created by Rancher Labs, it strips out cloud-provider-specific code and replaces etcd with SQLite, reducing the binary to under 100MB.

• Memory footprint: ~512MB RAM for the server node (vs 2GB+ for full K8s)
• Single binary: One ~60MB binary handles everything
• ARM native: First-class ARM64 and ARMv7 support
• Batteries included: Traefik ingress, CoreDNS, and Flannel CNI pre-configured
• Production ready: CNCF certified Kubernetes conformant

Hardware Setup for a Pi Cluster

A minimal cluster needs at least 2 Pis; 3-4 is ideal for learning high availability:

Installing K3s on the Master Node

Start with the server (master) node. Ensure each Pi has a unique hostname and static IP:

# Set hostname on the master node
sudo hostnamectl set-hostname k3s-master

# Assign a static IP (edit dhcpcd.conf or use NetworkManager)
# Example: 192.168.1.100 for master

# Enable cgroups (required for K3s)
# Edit /boot/firmware/cmdline.txt and add at the END of the line:
# cgroup_enable=cpuset cgroup_memory=1 cgroup_enable=memory
sudo nano /boot/firmware/cmdline.txt
sudo reboot

# Install K3s server
curl -sfL https://get.k3s.io | sh -

# Check the server status
sudo systemctl status k3s
sudo kubectl get nodes

# Get the node token (needed for workers)
sudo cat /var/lib/rancher/k3s/server/node-token

Joining Worker Nodes to the Cluster

# On each worker Pi, set a unique hostname:
sudo hostnamectl set-hostname k3s-worker-1  # or worker-2, worker-3

# Enable cgroups (same as master)
# Edit /boot/firmware/cmdline.txt, add:
# cgroup_enable=cpuset cgroup_memory=1 cgroup_enable=memory
sudo reboot

# Join the cluster (replace with your master IP and token)
curl -sfL https://get.k3s.io | K3S_URL=https://192.168.1.100:6443 
  K3S_TOKEN="your-node-token-here" sh -

# Back on the master, verify all nodes:
sudo kubectl get nodes
# NAME          STATUS   ROLES                  AGE   VERSION
# k3s-master    Ready    control-plane,master   10m   v1.30.x+k3s1
# k3s-worker-1  Ready    <none>                2m    v1.30.x+k3s1
# k3s-worker-2  Ready    <none>                1m    v1.30.x+k3s1

Deploying Your First Application

# Create a deployment YAML
cat << 'EOF' > nginx-deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
  labels:
    app: nginx
spec:
  replicas: 3
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx:alpine
        ports:
        - containerPort: 80
        resources:
          limits:
            memory: "128Mi"
            cpu: "250m"
---
apiVersion: v1
kind: Service
metadata:
  name: nginx-service
spec:
  type: NodePort
  selector:
    app: nginx
  ports:
    - port: 80
      targetPort: 80
      nodePort: 30080
EOF

# Deploy it
sudo kubectl apply -f nginx-deployment.yaml

# Check the deployment
sudo kubectl get pods -o wide
sudo kubectl get services

# Access at http://<any-node-ip>:30080

Monitoring with kubectl and Dashboard

# Install the Kubernetes Dashboard
GITHUB_URL=https://github.com/kubernetes/dashboard/releases
VERSION_KUBE_DASHBOARD=$(curl -w '%{url_effective}' -I -L -s -S 
  $GITHUB_URL/latest -o /dev/null | sed -e 's|.*/||')
sudo kubectl create -f 
  https://raw.githubusercontent.com/kubernetes/dashboard/${VERSION_KUBE_DASHBOARD}/aio/deploy/recommended.yaml

# Create admin user for dashboard
cat << 'EOF' | sudo kubectl apply -f -
apiVersion: v1
kind: ServiceAccount
metadata:
  name: admin-user
  namespace: kubernetes-dashboard
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: admin-user
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: cluster-admin
subjects:
- kind: ServiceAccount
  name: admin-user
  namespace: kubernetes-dashboard
EOF

# Get the login token
sudo kubectl -n kubernetes-dashboard create token admin-user

# Access dashboard via port-forward
sudo kubectl port-forward -n kubernetes-dashboard service/kubernetes-dashboard 8443:443 --address=0.0.0.0

Persistent Storage and Networking

K3s includes local-path-provisioner for persistent volumes:

# Local path storage class is the default in K3s
sudo kubectl get storageclass
# NAME                   PROVISIONER             AGE
# local-path (default)   rancher.io/local-path   30m

# For NFS shared storage across nodes:
sudo apt install nfs-kernel-server  # On one Pi
sudo mkdir -p /srv/nfs/k3s
echo "/srv/nfs/k3s *(rw,sync,no_subtree_check)" | sudo tee -a /etc/exports
sudo exportfs -ra

# Useful networking commands
sudo kubectl get ingress
sudo kubectl get endpoints
sudo kubectl describe service nginx-service

Frequently Asked Questions

{“@context”: “https://schema.org”, “@type”: “FAQPage”, “mainEntity”: [{“@type”: “Question”, “name”: “How many Raspberry Pis do I need for a K3s cluster?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “Minimum 2 (1 server + 1 worker), but 3 is recommended. For high availability, use 3 server nodes with an embedded etcd cluster. A 4-node cluster (1 server + 3 workers) is ideal for learning.”}}, {“@type”: “Question”, “name”: “Can I run K3s on Raspberry Pi 3?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “Technically yes, but not recommended. The Pi 3 has only 1GB RAM, which barely covers K3s server requirements. Worker nodes on Pi 3 can run a few lightweight pods. Pi 4 with 4GB+ is the practical minimum.”}}, {“@type”: “Question”, “name”: “What is the cost of building a Raspberry Pi Kubernetes cluster in India?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “A 3-node cluster costs approximately u20b915,000-25,000 depending on Pi model and accessories. This includes 3x Pi 4 (4GB), cases, SD cards, ethernet cables, and a switch. Significantly cheaper than cloud Kubernetes.”}}, {“@type”: “Question”, “name”: “K3s vs MicroK8s vs full Kubernetes for Raspberry Pi?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “K3s is the best choice for Pi clusters due to lowest resource usage. MicroK8s works well for single-node setups. Full Kubernetes (kubeadm) requires 2GB+ RAM per node and is overkill for Pi.”}}, {“@type”: “Question”, “name”: “Can I use WiFi instead of Ethernet for the cluster?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “WiFi works but is unreliable for Kubernetes. Cluster communication is latency-sensitive, and WiFi adds jitter. Always use Ethernet for the cluster network. WiFi can be used for external access.”}}]}