Kubernetes + Longhorn Lab Setup

Automated deployment of a complete Kubernetes cluster with storage and monitoring from your host machine.

Prerequisites

• 3 Ubuntu VMs with SSH access (password auth)
• Python 3.10+ on your host machine
• Extra disk on worker VMs for Longhorn storage (e.g., /dev/vdb)

Quick Start

# 1. Create virtual environment
uv venv .venv
source .venv/bin/activate
uv pip install -r requirements.txt

# 2. Edit config with your VM IPs
nano config.yaml

# 3. Run deployment
python deploy.py

What Gets Installed

Component	Version	Description
Kubernetes	1.35	Container orchestration
Calico	3.31.3	CNI networking
Longhorn	latest	Distributed block storage
Metrics Server	latest	Resource metrics
ingress-nginx	latest	Ingress controller
cert-manager	latest	Certificate management
Prometheus	latest	Metrics & alerting
Grafana	latest	Dashboards
Loki + Promtail	latest	Log aggregation

Configuration

Edit config.yaml:

ssh:
  user: ubuntu
  password: your-password

nodes:
  control:
    ip: 192.168.1.100
    hostname: k8s-control
  workers:
    - ip: 192.168.1.101
      hostname: k8s-worker-1
    - ip: 192.168.1.102
      hostname: k8s-worker-2

kubernetes:
  version: "1.35"
  pod_cidr: "10.244.0.0/16"
  service_cidr: "10.96.0.0/12"
  calico_version: "3.31.3"

longhorn:
  disk: /dev/vdb
  mount: /var/lib/longhorn

monitoring:
  grafana_password: admin123

Accessing Web UIs

After deployment, SSH to control plane and run the helper script:

ssh user@<control-plane-ip>
~/port-forward.sh

Then access from your browser:

Service	URL	Credentials
Grafana	http://<control-ip>:3000	admin / admin123
Prometheus	http://<control-ip>:9090	-
Alertmanager	http://<control-ip>:9093	-
Longhorn	http://<control-ip>:8080	-

Reset Cluster

To start fresh, reset all nodes:

# Control plane
ssh user@<control-ip> "echo 'password' | sudo -S bash -c 'kubeadm reset -f; rm -rf /etc/cni/net.d /var/lib/etcd ~/.kube'"

# Workers (includes Longhorn disk wipe)
ssh user@<worker-ip> "echo 'password' | sudo -S bash -c 'kubeadm reset -f; rm -rf /etc/cni/net.d; umount /var/lib/longhorn 2>/dev/null; wipefs -a /dev/vdb 2>/dev/null; rm -rf /var/lib/longhorn'"

Troubleshooting

# Check nodes
kubectl get nodes -o wide

# Check all pods
kubectl get pods -A

# Check specific namespace
kubectl get pods -n longhorn-system
kubectl get pods -n monitoring

# View logs
kubectl logs -n <namespace> <pod-name>

# Describe pod for events
kubectl describe pod -n <namespace> <pod-name>

Deployment Phases

1. Prepare nodes - Install containerd, kubelet, kubeadm
2. Init control plane - kubeadm init + Calico CNI
3. Join workers - kubeadm join + Longhorn disk prep
4. Install add-ons - Helm, Metrics Server, Longhorn, ingress-nginx, cert-manager
5. Install monitoring - Prometheus stack, Loki, Promtail
6. Final setup - Apps namespace, health check, port-forward script

Files

• deploy.py - Main deployment script
• config.yaml - Cluster configuration
• port-forward.sh - Helper script (also uploaded to control plane)