Longhorn Storage

Cloud-native distributed block storage for Kubernetes.

Overview

Longhorn provides replicated persistent volumes for your K8s cluster:

• 3 copies across nodes (HA)
• Incremental backups to S3
• Built-in snapshot + restore
• Simple UI for management

Why Longhorn?

Problem Solved

K8s pods are ephemeral — when they die, their data dies. Longhorn solves this:

Before: Pod dies → PVC data lost
After:  Pod dies → PVC survives, new pod reattaches

How It Works

flowchart LR
  subgraph Nodes
    n1["Node 1"]
    n2["Node 2"]
    n3["Node 3"]
  end
  
  pvc["PVC (10Gi)"]
  r1["Replica 1"]
  r2["Replica 2"]
  r3["Replica 3"]
  
  pvc --> r1
  pvc --> r2
  pvc --> r3
  r1 --> n1
  r2 --> n2
  r3 --> n3

Each PVC has 3 replicas on different nodes — if one node dies, data survives.

Install

Longhorn is installed as a system component (not in Apps/):

# Install via Helm
helm repo add longhorn https://charts.longhorn.io
helm install longhorn longhorn/longhorn --namespace longhorn-system --create-namespace

# Or via kubectl
kubectl apply -f https://raw.githubusercontent.com/longhorn/longhorn/v1.7.0/deploy/longhorn.yaml

Usage in Manifests

Example PVC

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: openclaw-data
spec:
  accessModes:
    - ReadWriteOnce
  storageClassName: longhorn
  resources:
    requests:
      storage: 10Gi

Key Settings

Access the UI

After install, port-forward or expose via ingress:

# Quick access (terminal)
kubectl -n longhorn-system port-forward svc/longhorn-frontend 8080:80

# Or via ingress (production)
https://longhorn.yourdomain.com

UI features:

• Volume management (create, delete, snapshot)
• Backup configuration (S3 credentials)
• Node health monitoring
• Engine image upgrades

Backup to S3

Configure S3 backup in Longhorn settings:

# In Longhorn UI: Settings → Backup
endpoint: s3.us-east-1.amazonaws.com
bucketName: longhorn-backups
accessKey: <your-key>
secretKey: <your-secret>

Then create scheduled backups:

# In Volume settings
Backup: Enabled
Schedule: 0 2 * * *  # Daily at 2 AM

Lessons Learned

What worked: Longhorn is production-ready — simple, reliable, great UI.

What I’d improve:

• Enable encryption at rest (LUKS)
• Set up S3 backup immediately (don’t wait for disaster)
• Tune replica count based on node count (3 for 3+ nodes, 2 for 2 nodes)

Related

• Windows VM — Uses Longhorn for 75Gi PVC
• Minecraft — Uses hostPath instead (contrast)