Project Nessie

"Project Nessie: Git‑Like Versioning for Lakehouse Tables and Catalogs"
Modern lakehouses demand more than fast query engines—they demand safe collaboration across teams and tools without duplicating data or fragmenting catalog state. This book targets experienced data platform engineers, architects, and senior practitioners who already run Iceberg-style tables in production and now need Git-like workflows for catalog changes: branching, isolation, promotion, and auditability, implemented with rigor rather than analogy.
You’ll build a precise mental model of what Nessie versions (catalog state, not files), then dive into references, commits, and the commit graph as operational primitives. From there, the book explains transactional semantics such as atomic multi-table commits, how engines interact with those guarantees, and where failures and concurrency hazards actually arise. Practical chapters show how to integrate Nessie as an Apache Iceberg catalog across multiple engines, debug with the REST/OpenAPI contract, and design dev/test/prod workflows using branches and tags. You’ll learn decision criteria for merge vs rebase vs cherry-pick in data promotion, plus time travel strategies for reproducible pipelines and audit-ready change reporting.
Coverage is production-minded: deployment patterns, observability, security boundaries, and safe garbage collection with retention economics and compatibility runbooks. The emphasis is on battle-tested architecture, operational guardrails, and techniques that scale with real teams and real write contention.