Re: MultiXact\SLRU buffers configuration

On Sat, Feb 19, 2022 at 6:34 PM Andrey Borodin <x4mmm(at)yandex-team(dot)ru> wrote:
> There's feature freeze approaching again. I see that you are working on moving SLRUs to buffer pools, but it is not clear to which PG version it will land. And there is 100% consensus that first patch is useful and helps to prevent big issues. Maybe let's commit 1'st step without lifting default xact_buffers limit? Or 1st patch as-is with any simple technique that prevents linear search in SLRU buffers.

Hi Andrey,

Yeah, the SLRU/buffer pool thing would be complete enough to propose
for 16 at the earliest. I posted the early prototype to see what sort
of reaction the concept would get before doing more work; I know
others have investigated this topic too... maybe it can encourage more
patches, experimental results, ideas to be shared... but this is not
relevant for 15.

Back to this patch: assuming we can settle on a good-enough-for-now
replacement algorithm, do we want to add this set of 7 GUCs? Does
anyone else want to weigh in on that? Concretely, this patch adds:

multixact_offsets_buffers
multixact_members_buffers
subtrans_buffers
notify_buffers
serial_buffers
xact_buffers
commit_ts_buffers

I guess the people at
https://ottertune.com/blog/postgresql-knobs-list/ would be happy if we
did. Hopefully we'd drop the settings in a future release once we
figure out the main buffer pool thing (or some other scheme to
automate sizing).