Skip to content

README.md

Latest commit

 

History

History
114 lines (78 loc) · 3.45 KB

README.md

File metadata and controls

114 lines (78 loc) · 3.45 KB

SBD Operator

A Kubernetes operator for managing STONITH Block Device (SBD) configurations and remediations for high-availability clustering. The operator provides automated node remediation when nodes become unresponsive by leveraging shared block storage for fencing operations.

Overview

The SBD operator implements a cloud-native approach to Storage-Based Death (SBD) for Kubernetes environments where traditional out-of-band management (IPMI, iDRAC) is unavailable. It uses shared block storage to provide reliable node fencing capabilities, ensuring data consistency and preventing split-brain scenarios in stateful workloads.

Architecture

The operator consists of two main components:

  • SBD Operator: Manages SBDConfig and SBDRemediation custom resources and deploys the SBD agent
  • SBD Agent: Runs as a DaemonSet on cluster nodes, handling local watchdog operations and shared storage communication

Key Features

  • Shared Storage Fencing: Uses CSI block PVs with concurrent multi-node access for inter-node communication
  • Dual Watchdog System: Combines shared storage watchdog with local kernel watchdog for robust failure detection
  • Kubernetes Integration: Native CRDs for configuration management and remediation requests
  • Prometheus Metrics: Built-in monitoring and observability
  • Split-Brain Prevention: Shared storage arbitration ensures cluster consistency

Custom Resources

SBDConfig

Defines the SBD configuration for the cluster:

  • Shared block device PVC name
  • Timeout settings
  • Watchdog device path
  • Node exclusion lists
  • Reboot methods

SBDRemediation

Triggers node remediation operations:

  • Target node specification
  • Remediation status tracking
  • Integration with Medik8s Node Healthcheck Operator

Quick Start

Prerequisites

  • Kubernetes cluster with CSI driver supporting volumeMode: Block
  • Shared block storage with concurrent multi-node access (e.g., Ceph RBD, cloud provider shared volumes)
  • Cluster nodes with kernel watchdog support

Installation

  1. Install the operator:
make deploy
  1. Create an SBDConfig:
kubectl apply -f config/samples/medik8s_v1alpha1_sbdconfig.yaml

Development

Build and test locally:

# Build the operator
make build

# Run tests
make test

# Run e2e tests
make test-e2e

# Build and push images
make docker-build docker-push IMG=<your-registry>/sbd-operator:tag

Documentation

Comprehensive documentation is available in the docs/ directory:

Testing

The project includes comprehensive testing:

  • Unit Tests: make test
  • E2E Tests: make test-e2e

E2E tests run against a deployed operator and verify functionality end-to-end.

Contributing

  1. Follow Go best practices and project coding standards
  2. Include comprehensive tests for new features
  3. Update documentation for user-facing changes
  4. Ensure all tests pass before submitting PRs

Requirements

  • Go 1.21+
  • Kubernetes 1.28+
  • Docker/Podman for container builds
  • Make for build automation

License

Licensed under the Apache License 2.0. See LICENSE for details.

Support

For issues, questions, or contributions, please use the GitHub issue tracker.