On Wed, Feb 4, 2026 at 10:10:55AM -0500, Tom Lane wrote:
> Heikki Linnakangas <hlinnaka@iki.fi> writes:
> > On 03/02/2026 18:33, Tom Lane wrote:
> >> Apparently, because its file system sucks.
>
> > I bet it's the GCC compile farm's storage system that sucks.
>
> Certainly possible; I have no visibility into what's actually
> in there. But it seems to have about a dozen drives, which
> ought to give a decent amount of bandwidth ...
>
> > Buildfarm
> > animal 'douc' doesn't seem to have that problem:
> > https://buildfarm.postgresql.org/cgi-bin/show_history.pl?nm=douc&br=REL_16_STABLE
>
> douc isn't running the TAP tests, so it's not directly comparable.
> But I don't find that a 25-minute BF cycle time without TAP tests
> is impressive.
>
> Anyway, I would be happy if someone would put forward evidence
> refuting my results, and even more happy if community members
> trying to work on this patch had access to AIX hardware that
> wasn't ridiculously slow. As things stand right now, I have
> zero motivation to continue working on it.
I would have given up on this AIX effort long ago. It was clear to me
the developer work wasn't worth the value to the user base, and the
quality of the patches has been consistently low, as pointed out by
Robert Haas recently:
https://www.postgresql.org/message-id/CA+Tgmobh=69ud8J62XFAErS-R=mxXC_9RT85hoFMPG_M8+650Q@mail.gmail.com
Patch authors have been enthusiastic, but patch activity has been
inconsistent. The recent testing problems is just one in a long list of
hurdles to resurrect this platform.
Probably the only value to the community of adding AIX is platform
diversity testing. In hindsight, a verbal meeting two years ago might
have avoided many of these problems, though patch author turnover might
have made that ineffective.
I think the gap between AIX features and other platforms, the amount of
work required to make AIX a _clean_ port, the amount of time and skills
patch authors have, and the value the community of having an AIX port is
just too wide, though I will say I am surprised the patch authors have
gotten as far as they have.
--
Bruce Momjian <bruce@momjian.us> https://momjian.us
EDB https://enterprisedb.com
Do not let urgent matters crowd out time for investment in the future.
