Re: Load distributed checkpoint

From: "Jim C(dot) Nasby" <jim(at)nasby(dot)net>
To: Bruce Momjian <bruce(at)momjian(dot)us>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Heikki Linnakangas <heikki(at)enterprisedb(dot)com>, ITAGAKI Takahiro <itagaki(dot)takahiro(at)oss(dot)ntt(dot)co(dot)jp>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Load distributed checkpoint
Date: 2006-12-30 04:08:04
Message-ID: 20061230040804.GE71246@nasby.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers pgsql-patches

On Fri, Dec 29, 2006 at 09:02:11PM -0500, Bruce Momjian wrote:
> Tom Lane wrote:
> > "Jim C. Nasby" <jim(at)nasby(dot)net> writes:
> > > I believe there's something similar for OS X as well. The question is:
> > > would it be better to do that, or to just delay calling fsync until the
> > > OS has had a chance to write things out.
> >
> > A delay is not going to help unless you can suppress additional writes
> > to the file, which I don't think you can unless there's very little
> > going on in the database --- dirty buffers have to get written to make
> > room for other pages, checkpoint in progress or no.
>
> I am afraid a delay between write and fsync is the only portable option
> we have right now --- there is hope that since the check point write, we
> will not have a huge number of dirty buffers at the start of the
> checkpoint that need to be written out.

We could potentially control that behavior, too. For example, trying to
suppress writes to a given file when we're getting ready to checkpoint
it (of course, separating checkpointing into a file-by-file operation is
a non-trivial change, but it should be possible).
--
Jim Nasby jim(at)nasby(dot)net
EnterpriseDB http://enterprisedb.com 512.569.9461 (cell)

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Bruce Momjian 2006-12-30 04:29:11 Re: TODO: GNU TLS
Previous Message Joshua D. Drake 2006-12-30 02:30:21 Re: psql possible TODO

Browse pgsql-patches by date

  From Date Subject
Next Message Jim C. Nasby 2006-12-30 04:46:27 Re: column ordering, was Re: [PATCHES] Enums patch v2
Previous Message Bruce Momjian 2006-12-30 02:02:11 Re: Load distributed checkpoint