SREcon26 Americas - 5 Wrong Hypotheses about PostgreSQL Multi-Transaction Locks @UsenixOrg
SREcon26 Americas - 5 Wrong Hypotheses about PostgreSQL Multi-Transaction Locks  @UsenixOrg
Uploaded April 2026 | Updated September 2026, 3 weeks ago
5 Wrong Hypotheses about PostgreSQL Multi-Transaction Locks

Clint Byrum, HashiCorp, an IBM Company

Working on complex systems and expecting to be right are not always compatible things. PostgreSQL is an incredibly complex component of complex systems, and how we use it is just as important as how it works internally.

We have spent several years operating a very busy AWS Aurora PostgreSQL database with varying levels of reliability, and we got pretty good at finding ways to break it with MultiXacts. This meant incidents, experiments, and learning to get a little less wrong with each iteration. Join us to learn about two very important things:

1. Several, decreasingly wrong ideas about how MultiXact locks work in PostgreSQL with data and analysis from real incidents.

2. How to persevere through the stress of incidents and keep learning in a complex system despite knowing that you're at least a little wrong, all the time.

View the full SREcon26 Americas program at usenix.org/conference/srecon26americas/program
SREcon26 Americas - 5 Wrong Hypotheses about PostgreSQL Multi-Transaction LocksSREcon25 Europe/Middle East/Africa - Utilization Is the Key to Efficiency: What It Takes to Run...NSDI 26 - PolicyCache: Intra-flow Learning in Congestion ControlNSDI 26 - From Source to Solution: Tackling Packet Losses in Large-scale Cloud Gaming...
USENIX |

SREcon26 Americas - 5 Wrong Hypotheses about PostgreSQL Multi-Transaction Locks

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER