Re: Horrible CREATE DATABASE Performance in High Sierra

Hi,

On 2017-10-02 18:33:17 -0400, Tom Lane wrote:
> Andres Freund <andres(at)anarazel(dot)de> writes:
> > To demonstrate what I'm observing here, on linux with a fairly fast ssd:
> > ...
>
> I tried to replicate this test as closely as I could on the Mac hardware
> I have laying about.

Thanks!

> I only bothered with the synchronous_commit=off case, though, since
> you say that shows the worst effects.

That's the case on linux at least, but I'd suspect it's a much more
general thing - you just can't get that much data dirty with pgbench
otherwise.

> I'm kind of surprised that machine B doesn't show obvious tanking in this
> test given that msync() makes it suck so badly at copying a database.
> I wonder what is different from the kernel's standpoint ... maybe the
> sheer number of different files mmap'd by a single process during the
> copy?

Yea, that's curious. I've really no clue about OSX, so pardon my
question: With HEAD and CREATE DATABASE, is it IO blocked or kernel cpu
blocked?

> If we could arrange to not use pg_flush_after in copydir.c on macOS,
> I'd be okay with leaving it alone for the configurable flush_after
> calls. But I can't think of a way to do that that wouldn't be a
> complete kluge. I don't much want to do
>
> +#ifndef __darwin__
> pg_flush_data(dstfd, offset, nbytes);
> +#endif
>
> but I don't see any better alternative ...

What we could do is introduce a guc (~create_database_flush_data) that
controls whether we flush here, and have the defaults set differently on
OSX. Not sure if that's actually any better.

- Andres