DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com;
	h=received:message-id:date:from:reply-to:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references;
	b=fzdyxgwM0EHwlpwkiVXl4cVEn56OlccDYj6X+HmyIRR4ok1+k3IEIf1Z/zrq7qLN5vX1jkLAToOdc8qFKZAgduwmTHpZx6+WMJVvYZ1By4a1WAXT0OwexG6ZWrDqWtyyevp9yEXodPGFFUW8dq7dwCotazZm+P3MnyQhxUyS36E=
Message-ID: <59d991c405071811457ccdd3f1@mail.gmail.com>
Date: Mon, 18 Jul 2005 14:45:35 -0400
From: Christopher Petrilli <petrilli@gmail.com>
Reply-To: Christopher Petrilli <petrilli@gmail.com>
To: Vivek Khera <vivek@khera.org>
Subject: Re: Impact of checkpoint_segments under continual load conditions
Cc: pgsql-performance@postgresql.org
In-Reply-To: <A50B8A1F-71A1-4F63-A5F4-1AA58F89557D@khera.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
References: <59d991c4050717100874c08a86@mail.gmail.com>
	<A50B8A1F-71A1-4F63-A5F4-1AA58F89557D@khera.org>

On 7/18/05, Vivek Khera <vivek@khera.org> wrote:
>=20
> On Jul 17, 2005, at 1:08 PM, Christopher Petrilli wrote:
>=20
> > Normally, checkpoint_segments can help absorb some of that, but my
> > experience is that if I crank the number up, it simply delays the
> > impact, and when it occurs, it takes a VERY long time (minutes) to
> > clear.
>=20
> There comes a point where your only recourse is to throw hardware at
> the problem.  I would suspect that getting faster disks and splitting
> the checkpoint log to its own RAID partition would help you here.
> Adding more RAM while you're at it always does wonders for me :-)

My concern is less with absolute performance, than with the nosedive
it goes into.  I published some of my earlier findings and comparisons
on my blog, but there's a graph here:

http://blog.amber.org/diagrams/comparison_mysql_pgsql.png

Notice the VERY steep drop off.  I'm still trying to get rid of it,
but honestly, am not smart enough to know where it's originating.  I
have no desire to ever use MySQL, but it is a reference point, and
since I don't particularly need transactional integrity, a valid
comparison.

Chris
--=20
| Christopher Petrilli
| petrilli@gmail.com