Re: Why our Valgrind reports suck

=?utf-8?Q?=C3=81lvaro?= Herrera <alvherre(at)kurilemu(dot)de> writes:
> On 2025-May-11, Tom Lane wrote:
>> In particular, I had not realized that autovacuum
>> leaks a nontrivial amount of memory per relation processed (cf 0009),
>> and apparently has done for a few releases now. This is horrid in
>> databases with many tables, and I'm surprised that we've not gotten
>> complaints about it.

> In PGConf Germany we got a lightning talk[1] that reported a problem
> that might be explained by this: with 10 million of relations, the OOM
> killer gets invoked on autovacuum workers on the reported case, so
> essentially autovacuum doesn't work at all. So clearly there is somebody
> that would appreciate that this problem is fixed.

Interesting.

> He actually blames it on relcache, but who knows how correct that is.

I would not be surprised if the relcache is bloated too, but Valgrind
wouldn't think that's a leak. I wonder if it'd be worth setting up
a mechanism for autovacuum to drop the relcache entry for a table
once it's done with it.

regards, tom lane