Distributed Postgres with Citus
summary
Citus turns Postgres into a sharded, multi node database by adding a Citus extension to standard Postgres servers and a coordinator that rewrites queries into per shard SQL. It shards by a hash of the shard key, runs aggregates in parallel, and uses two‑phase commit for distributed transactions plus a cluster wide deadlock detector. Backup and recovery use WAL shipping with wal‑e / wal‑g and shards can be rebalanced to add capacity or enable masterless coordinators ; cloud failovers take about 30s–60s.
AI-written summary · about
chapters (14)
- 0:00Introduction to speaker
- 0:37Why Postgres is popular
- 1:44Scaling challenges
- 3:01Citus architecture overview
- 5:01Citus internals and sharding
- 7:22Scaling SQL complexities
- 8:07Postgres replication methods
- 12:58Disaster recovery techniques
- 15:40Distributing PostgreSQL
- 18:04Distributed query planning
- 22:40Distributed transactions
- 24:59Concurrency and deadlocks
- 28:27Distributed deadlock detection
- 32:38Q&A and conclusion
related talks
08mFoundations of Distributed PostgreSQL with Citus
31mSharding Without the Complexity
31mKEYNOTE: The Distributed PostgreSQL Problem & How Citus Solves it
33mCitus & Patroni: The Key to Scalable and Fault-Tolerant PostgreSQL
28mEfficiently distributing Postgres with Citus – How to choose the right shard key?
32m8 keys to the growth of Citus & Distributed PostgreSQL
28mCitus on Kubernetes
16mKeynote – Umur Cubukcu
