Re: Vacuum ERRORs out considering freezing dead tuples from before OldestXmin

Hi,

On 2024-07-21 12:51:51 -0400, Tom Lane wrote:
> Melanie Plageman <melanieplageman(at)gmail(dot)com> writes:
> > When I run it on my machine with some added logging, the space taken
> > by dead items is about 330 kB more than maintenance_work_mem (which is
> > set to 1 MB). I could roughly double the excess by increasing the
> > number of inserted tuples from 400000 to 600000. I'll do this.

> mamba, gull, and mereswine are 32-bit machines, which aside from
> being old and slow suffer an immediate 2x size-of-test penalty:

I think what we ought to do here is to lower the lower limit for memory usage
for vacuum. With the new state in 17+ it basically has become impossible to
test multi-pass vacuums in a way that won't get your test thrown out - that's
bad.

> I do not think the answer to this is to nag the respective animal owners to
> raise PG_TEST_TIMEOUT_DEFAULT. IMV this test is simply not worth the cycles
> it takes, at least not for these machines.

This specific area of the code has a *long* history of bugs, I'd be very loath
to give up testing.

Greetings,

Andres Freund