Re: Initdb-time block size specification

On Tue, Sep 5, 2023 at 9:04 AM Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>
> On Sat, Sep 2, 2023 at 3:09 PM Tomas Vondra
> <tomas(dot)vondra(at)enterprisedb(dot)com> wrote:
> > Except that no one is forcing you to actually go 130mph or 32mph, right?
> > You make it seem like this patch forces people to use some other page
> > size, but that's clearly not what it's doing - it gives you the option
> > to use smaller or larger block, if you chose to. Just like increasing
> > the speed limit to 130mph doesn't mean you can't keep going 65mph.
> >
> > The thing is - we *already* allow using different block size, except
> > that you have to do custom build. This just makes it easier.
>
> Right. Which is worth doing if it doesn't hurt performance and is
> likely to be useful to a lot of people, and is not worth doing if it
> will hurt performance and be useful to relatively few people. My bet
> is on the latter.

Agreed that this doesn't make sense if there are major performance
regressions, however the goal here is patch evaluation to measure this
against other workloads and see if this is the case; from my localized
testing things were within acceptable noise levels with the latest
version.

I agree with Tomas' earlier thoughts: we already allow different block
sizes, and if there are baked-in algorithmic assumptions about block
size (which there probably are), then identifying those or places in
the code where we need additional work or test coverage will only
improve things overall for those non-standard block sizes.

Best,

David