Linux 7.0 Cuts PostgreSQL Performance in Half

https://winbuzzer.com/2026/04/05/aws-engineer-reports-postgresql-perf-halved-by-linux-7-kernel-change-xcxwbn/

120 points · 12 comments · view on lemmy.world

12 Comments

Endmaker@ani.social · 63 pts · 152d (2 replies)

kernel maintainers are pushing the fix burden onto PostgreSQL

Maybe it isn't applicable in this context, but didn't Linus Torvalds send an angry email on an adjacent topic, but regarding the same philosophy?

Found it: we do not break userspace

Disclaimer: I am a noob when it comes to Linux and building operating systems.

Laser@feddit.org · 23 pts · 152d

Well, technically, it's not broken, just slower

BlackEco@lemmy.blackeco.com · 8 pts · 152d

Regardless of Linus's philosophy, such a change in Postgresql would need to be backported to previous versions, because migrating from one major release to another is far from being straightforward.

rimu@piefed.social · 31 pts · 152d (2 replies)

Wouldn't it make sense to merge in rseq support well in advance of removing PREEMPT_NONE, not do both at the same time?

deegeese@sopuli.xyz · 7 pts · 151d

But that hurts the ego of the kernel developer because they have to wait for the next release.

Die4Ever@retrolemmy.com · 2 pts · 151d

sounds like rseq has been around since 2018 https://www.phoronix.com/news/Restartable-Sequences-Speed

realitaetsverlust@piefed.zip · 18 pts · 151d (2 replies)

I think it's a bit weird that postgres hasn't done any testing on the new kernel. That's something I would kinda expect from a major database. The fact this was only found in production is a bit ... weird.

balian@lemmy.libertarianfellowship.org · 1 pts · 149d (1 reply)

Why would they? The current stable kernel release is still v6.19

realitaetsverlust@piefed.zip · 1 pts · 149d

Because that's what you do with important software? You don't want to stack up potential issues until the final release for the public.

just_another_person@lemmy.world · 9 pts · 152d

This has been known for awhile, and this was already accepted to be an issue with PH memory handling. Not weird, rare or otherwise, it will get fixed.

Die4Ever@retrolemmy.com · 6 pts · 151d

I'll be curious to see benchmarks of PREEMPT_NONE vs rseq once PostgreSQL patches this

Which is expected to be faster in the end?

obbeel@lemmy.eco.br · 0 pts · 151d

Maybe Linux is considering prioritizing consumer hardware (safety, less room for human error)? Android is already a thing, so maybe there is pressure for prioritizing consumer hardware?