High Availability with PostgreSQL and Pacemaker

on YouTube ↗55m

" summary

Claims a way to lose no committed transactions by moving replication below PostgreSQL using DRBD block-level synchronous writes and keeping Postgres on an LVM/XFS volume, with Pacemaker-managed VIP and service orchestration to detect failure, promote the replica and start services. Examples show a two-node build from scratch, corosync for cluster messaging, Postgres 9.1 and pgbench for workload, automatic resync on rejoin, and observed live failovers around 10–20 second failover under VMs; useful caveats on barriers and stickiness.

AI-written summary · about

" chapters (12)

1. 0:00Start
2. 4:15Log Shipping
3. 12:52Build a Cluster
4. 30:10Set Up Lvm
5. 30:23Linux Volume Manager
6. 31:02Create a Volume
7. 31:24Create a Volume Group
8. 41:15Configure Koro Sync
9. 45:53Quorum Policy
10. 50:37Configure a Master-Slave Resource
11. 52:43Setting Up the Postgres Resource Monitor
12. 53:33Failover

" related talks

• 50mEvolution Of PostgreSQL Replication and High Availability (upto ~ 99.999%)Stacy Scoggins · PGConf NYC 2021
• 54mBreakout SessionDavid Pacheco · PGConf Silicon Valley 2015
• 36mDesigning Layered High‑Availability Architectures for PostgreSQL on a BudgetUmair Shahid · SREcon26 Americas
• 50mMoving from MSSQL to pgSQL – High Availability and Disaster RecoveryRick Lowe · Postgres Conference Orlando 2025
• 41mEnsuring High Availability in Self-Managed PostgreSQL with PatroniArnab Saha · Postgres Conference Orlando 2025
• 57mReliability features in Postgres ProLeonid Albrekht · PGConf India 2025
• 49mPostgreSQL Replication - The Most Exciting Technology on EarthEdward Snajder · Open Source Bridge 2013
• 47mPostgreSQL High Availability: The Considerations and the CandidatesHamid Quddus Akhtar · Percona Live 2022