Cluster Automation

MariaDB Galera Cluster Automation & Multi-Master Sync
A production-grade knowledge base for running active-active MariaDB at scale.

🌐 Visit the site → clusterautomation.org

---

What this is

clusterautomation.org is a deep, production-focused reference for the engineers who keep MariaDB Galera clusters healthy — database administrators, DevOps and platform teams, and anyone automating multi-master replication. It goes past "hello world" tutorials into the mechanics that actually decide whether an active-active cluster stays certified under load: write-set certification, flow control, quorum arithmetic, state transfer, failover, and backup/restore.

Every page is written for real deployments on MariaDB 10.6–11.x with Galera 4 — explicit configuration, copy-ready Python and Bash automation, and version-specific caveats rather than toy examples.

What's inside

49 in-depth guides — 150,000+ words — organised into three areas, each with a custom, hand-authored diagram and structured data:

• Core Architecture & Fundamentals — how Galera works under the hood: the wsrep API, synchronous certification-based replication, write-set conflict resolution, multi-master topologies, quorum & automated failover, network security, Python monitoring & alerting, and read-only fallback routing.
• Cluster Setup & Node Management — operating a production cluster: hardware sizing, wsrep.cnf tuning and validation, bootstrapping, state-transfer methods, graceful node join/leave, health monitoring, and startup error recovery.
• Backup & Restore Sync — protecting and rebuilding cluster state: non-blocking mariabackup streaming SST, gcache sizing for fast IST recovery, point-in-time recovery in multi-master, and consistent restore drills.

Why it's different

• Depth over breadth. Each guide is a complete, self-contained procedure with runnable code, parameter references, verification steps, and troubleshooting — not a stub.
• Original diagrams. Every architecture, sequence, and state-transition diagram is a hand-authored inline SVG that adapts to light/dark themes — no stock images.
• Accessible & fast. Built to WCAG 2 AA (verified with axe-core) and a mobile Lighthouse performance budget, with valid JSON-LD structured data on every page.
• Offline-ready. A fully installable PWA — the whole reference works without a connection.
• Self-contained. No third-party scripts or CDNs; the only outbound links are curated references inside the content.

Built with

• Eleventy 3 — static site generator (Nunjucks layouts, Markdown content).
• markdown-it — anchors, task-lists, KaTeX math, PrismJS syntax highlighting.
• Progressive Web App — web manifest + service worker, installable and offline-capable.
• Cloudflare Workers — static-asset hosting at the edge.

Local development

npm install        # install dependencies
npm run build      # validate code blocks + generate icons + build into _site/
npm run serve      # local dev server with live reload
npm run deploy     # build, then deploy to Cloudflare

Project layout

content/                 Markdown guides (YAML front matter: title, description, type, dates…)
  content.11tydata.js    Shared data: permalink + short <title> for every page
src/
  _includes/             base.njk, doc.njk, partials (logo, icon sprite)
  _data/site.js          Site metadata, owner, and the content sections
  assets/                css, js, icon source + generated icons
  manifest.webmanifest   PWA manifest
  sw.js                  Service worker (network-first pages, stale-while-revalidate assets)
scripts/                 Icon generation + code-block validation
.eleventy.js             Build config (permalinks, markdown, passthrough, nav filters)

URLs mirror the content tree with the content/ prefix stripped, e.g. content/backup-and-restore-sync/index.md → /backup-and-restore-sync/.

License

Content © Cluster Automation. All rights reserved.

---

Explore the full reference at clusterautomation.org →