Skip to content

Longhorn Storage

Continuous persistent storage for Kubernetes workloads is provided by Longhorn.

Namespace: longhorn-system
Chart: longhorn/longhorn

1. Overview

Highly available block storage is provided for all stateful applications within the cluster. Reliability is ensured through data replication across nodes.

2. Architecture and Replication

  • Replication Factor: 1
  • Backup Target: Oracle Bucket S3 (Configured).
  • Frontend: Web UI accessible via internal cluster ingress.

3. Implementation Steps

  1. System Preparation: Ensure open-iscsi and nfs-common are installed on all worker nodes.
  2. Deployment: Installed via ArgoCD using the official Helm chart.
  3. StorageClass: Set as the cluster default StorageClass for automatic provisioning.

4. Maintenance and Backup

  1. Snapshot Schedule: Automated daily snapshots are performed for all active volumes.
  2. Node Maintenance: Nodes are drained prior to maintenance; volume replicas are automatically rebalanced by the scheduler.
  3. Health Monitoring: Volume health is monitored through the Longhorn UI and Prometheus alerts.