CockroachDB The Definitive Guide. Distributed Data at Scale (Guy Harrison, Jesse Seldess, Ben Darnell etc.)（Z-Library）

【One-Line Pitch】 A hands-on guide to building and running applications on CockroachDB, the distributed SQL database that trades a little latency for always-on availability and strict consistency. Best for developers, architects, and SRE/DevOps teams who already know relational databases and now need to operate at global scale. 【Book Arc】 - **Opening (~0%–10%)**: Frames CockroachDB against the history of databases—from monolithic RDBMSs to eventually-consistent NoSQL—and explains why the authors built a strongly consistent, highly available SQL system. Solves the "why does this database exist?" question. - **Early (~10%–35%)**: Moves into internals and setup: the KV storage layer (Pebble, LSM trees, SSTables), cluster architecture, installation on Linux/macOS/Windows, connection URLs, and creating clusters via Terraform. Solves "how do I get a cluster running and understand what's underneath?" - **Middle (~35%–55%)**: Core SQL and schema work—SELECT, data types (including ARRAY, JSONB, ENUM, VECTOR), views and materialized views, UDFs, UPSERT, and the beginnings of table/index creation. Solves "how do I model and query data?" - **Late (~55%–80%)**: Application design and implementation—CRUD operations, connection pools, prepared/parameterized statements, plus index strategy (composite, covering, inverted, partial, hash-sharded) and physical design. Solves "how do I make this fast and correct in real code?" - **Ending (~80%–100%)**: Operational maturity—migration strategies, multiregion resiliency, security, monitoring, and performance tuning. Solves "how do I keep this running at scale?" (Excerpts do not cover these chapters in detail.) 【Key Takeaways】 - **CockroachDB prioritizes consistency over availability, but still targets very high availability** (Early): It was built as a corrective to NoSQL's weak guarantees, keeping SQL and ACID transactions while distributing across nodes. - **It is wire-compatible with PostgreSQL** (Middle): Most Postgres drivers work unchanged—there are no CockroachDB-specific drivers—so existing Java, Go, Python, and JavaScript tooling transfers directly. - **The storage layer is an LSM-tree KV engine called Pebble** (Early): MemTables flush to SSTables across levels L0–L6, giving fast writes and efficient random reads; older versions used RocksDB. - **Schema and index design are first-class concerns** (Late): Composite, covering, inverted, partial, expression, full-text, spatial, and hash-sharded indexes each solve distinct access patterns, and the book treats index overhead as a real cost. - **Rich data types reduce the need for workarounds** (Middle): ARRAY, JSONB, ENUM, INET, UUID, and VECTOR (for embeddings) are native, alongside standard numeric, string, and temporal types. - **Views and functions are part of the modeling toolkit** (Middle): Standard and materialized views, plus SQL/PL-pgSQL user-defined functions, let you encapsulate logic in the database. - **Operational concerns are treated as engineering, not afterthoughts** (Late): Connection pooling, prepared statements, Terraform-based cluster provisioning, and the DB Console for metrics all appear as practical guidance. - **Real-world scale is the framing throughout** (Early): Case material such as Route processing billions of orders illustrates the elastic scale-up/scale-down promise with zero downtime. 【Reading Tips】 - **Skim the history chapter if you already know the NoSQL-vs-SQL story**; deep-read the storage-layer and architecture sections, since they explain *why* CockroachDB behaves the way it does under load. - **Treat the SQL and data-type chapters as a reference**, not a linear read—jump to the type or statement you need, then return to index design once your queries are stable. - **Pay closest attention to the index and application-design chapters**; this is where most real performance and correctness wins (or losses) happen. - **If you're migrating an existing Postgres app**, read the compatibility notes carefully—CockroachDB diverges in specific areas (e.g., no PostgreSQL XML functions). - **Keep the operational chapters for last**, but don't skip them: multiregion resiliency, security, and monitoring are where distributed databases actually get hard. 【Coverage Limits】 This guide is synthesized from stratified excerpts covering roughly the first half of the book; the later chapters on migration, multiregion resiliency, security, monitoring, and performance tuning are referenced but not detailed here.