| From: | Bruce Momjian <bruce(at)momjian(dot)us> |
|---|---|
| To: | Jon Nelson <jnelson+pgsql(at)jamponi(dot)net> |
| Cc: | Hannu Krosing <hannu(at)2ndquadrant(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>, Thom Brown <thom(at)linux(dot)com>, Andres Freund <andres(at)2ndquadrant(dot)com> |
| Subject: | Re: 9.4 regression |
| Date: | 2013-08-08 21:46:18 |
| Message-ID: | 20130808214618.GC23572@momjian.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Thu, Aug 8, 2013 at 04:33:05PM -0500, Jon Nelson wrote:
> > How much slower would it be if we wrote it with zeros after
> > posix_fallocate() --- that would still give use single extents. Has
> > anyone tested to see if the write without test_fallocate() still gives
> > us one extent?
>
> Actually, I did that test - I found that there was no longer a
> performance drop and possibly a slight performance benefit.
> I couldn't quite parse your last sentence, I assume you mean "does
> skipping the posix_fallocate altogether still produce one extent?"
> Sadly, I can't really test that properly as the ext4 filesystem I'm
> testing on is quite fresh and has lots and lots of free space.
Yes, that was the question. I meant to say:
Has anyone tested to see if the write without posix_fallocate() still
creates one extent per 16MB file, meaning the posix_fallocate()
is unnecessary?
--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ It's impossible for everything to be true. +
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Jon Nelson | 2013-08-08 21:46:46 | Re: 9.4 regression |
| Previous Message | Tom Lane | 2013-08-08 21:42:02 | Re: 9.4 regression |