Re: Potential "AIO / io workers" inter-worker locking issue in PG18?

From: Andres Freund <andres(at)anarazel(dot)de>
To: Marco Boeringa <marco(at)boeringa(dot)demon(dot)nl>
Cc: pgsql-bugs(at)lists(dot)postgresql(dot)org, Thom Brown <thom(at)linux(dot)com>
Subject: Re: Potential "AIO / io workers" inter-worker locking issue in PG18?
Date: 2025-10-07 19:56:34
Message-ID: tfeg5bh67smb3mv47ti47v7k2ekaegfhmt7ezwg4hv75pt5ljj@bbjsapryvagj
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

Hi,

On 2025-10-07 21:34:00 +0200, Marco Boeringa wrote:
> I have now been able to successfully run 'perf' after a VM restart and
> booting with the linux 6.8 kernel.

Unfortunately the profile is lacking most symbols, so it's not particularly
informative :(. You probably need to install the debug symbols package for
postgres (and probably libc)...

> Andres: I posted separately to you I had a suspicion a recent linux kernel
> update (6.8 --> 6.14) might be involved that coincided with the PG18/PostGIS
> 3.6.0 upgrade when I initially had multiple successful runs of my tool
> without the temporary stalling / delay in returning of postgres after
> booting to the older 6.8 kernel. This has now proven wrong: these results
> are of the 6.8 kernel runs, so the kernel is not involved, and I also see
> the issue with the older kernel.

Good to know.

Greetings,

Andres Freund

In response to

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message Marco Boeringa 2025-10-07 20:04:03 Re: Potential "AIO / io workers" inter-worker locking issue in PG18?
Previous Message Marco Boeringa 2025-10-07 19:49:14 Re: Potential "AIO / io workers" inter-worker locking issue in PG18?