The journey towards active replication in PostgreSQL
summary
Progress on PostgreSQL’s logical replication is mapped to concrete features and upgrades: logical decoding and replication slots (9.4), pgoutput and publication/subscription (10), truncate and replica identity improvements, commit-timestamp support, two‑phase transaction replication, and in v16 origin filtering to avoid replication loops enabling basic bi‑directional setups. Standby nodes can now publish via logical decoding from standby, non-superusers can create subscriptions, and streaming large transactions yields ~25–40% apply speedups; initial sync can use binary format between v16 nodes.
AI-written summary · about
chapters (11)
- 0:00Intro
- 2:16What is replication?
- 8:23Active-standby deployment model
- 16:08Active-active advantage and tradeoffs
- 18:32What does PostgreSQL need to support active-active?
- 20:16What does PostgreSQL need to better support active-active?
- 23:42Evolution of logical replication in PostgreSQL
- 29:21Allow logical decoding from standby
- 30:34Logical replication from standby
- 36:41Allow logical replication to copy tables in binary format
- 40:29Features to support active-active deployments
related talks
44mThe journey towards active-active replication in PostgreSQL
31mPast, Present, and Future: Logical Decoding and Replication in PostgreSQL
59mLogical Replication...LIVE!
42mlesser known Superpower of Postgres Logical Replication
01h 41mPostgreSQL Upgrades with Logical Replication & Minimal Downtime
48mPostgreSQL Replication: 20 Pitfalls and Solutions
48mReplication Refresh
54mUnderstanding Logical Decoding and Replication