DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.co.uk;
	h=X-YMail-OSG:Received:X-Mailer:Date:From:Reply-To:Subject:To:MIME-Version:Content-Type:Content-Transfer-Encoding:Message-ID;
	b=ghkiCdEvPgN67PyXVxZJSnzpEwYuCeACXT8+n6AhMfak1e7gtsBiJW/2fk+1DxUxbxYSm2N9eCREaHAHHbtVPtz9r2bt/xuvbbC5dM6rZ8MuP9qt32bVZ1Xhtos4LlDRdFhfTaeJnBMq2R4WzFFdwEDgfiNM2LWHLcNBMx5ymBo=;
Date: Wed, 14 May 2008 10:00:39 +0000 (GMT)
From: Valentin Bogdanov <valiouk@yahoo.co.uk>
Reply-To: valiouk@yahoo.co.uk
Subject: postgres overall performance seems to degrade when large SELECT are
	requested
To: pgsql-performance@postgresql.org
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Message-ID: <942836.27599.qm@web25805.mail.ukl.yahoo.com>

HI,

I have an application that maintains 150 open connections to a Postgres DB =
server. The application works fine without a problem for the most time.=20

The problem seem to arise when a SELECT that returns a lot of rows is execu=
ted or the SELECT is run on a large object. These selects are run from time=
 to time by a separate process whose purpose is to generate reports from th=
e db data.

The problem is that when the SELECTs are run the main application starts ru=
nning out of available connections which means that postgres is not returni=
ng the query results fast enough. What I find a little bit starnge is that =
the report engine's SELECTs operate on a different set of tables than the o=
nes the main application is using. Also the db box is hardly breaking a swe=
at, CPU and memory utilization are ridiculously low and IOwaits are typical=
ly less than 10%.

Has anyone experienced this? Are there any settings I can change to improve=
 throughput?  Any help will be greatly appreciated.


Thanks,
val=0A=0A=0A      _________________________________________________________=
_=0ASent from Yahoo! Mail.=0AA Smarter Email http://uk.docs.yahoo.com/nowyo=
ucan.html