X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [64.117.224.130]) by svr1.postgresql.org (Postfix) with ESMTP id 72797D1B550 for ; Thu, 11 Sep 2003 19:13:32 +0000 (GMT) Received: from svr1.postgresql.org ([64.117.224.193]) by localhost (neptune.hub.org [64.117.224.130]) (amavisd-new, port 10024) with ESMTP id 91400-08 for ; Thu, 11 Sep 2003 16:13:02 -0300 (ADT) Received: from mx1.neteconomist.com (mx1.neteconomist.com [65.122.18.6]) by svr1.postgresql.org (Postfix) with ESMTP id 5B38DD1B56B for ; Thu, 11 Sep 2003 16:13:00 -0300 (ADT) Received: from mx1.neteconomist.com (localhost.localdomain [127.0.0.1]) by mx1.neteconomist.com (8.12.8/8.12.8) with ESMTP id h8BJAJWg016439; Thu, 11 Sep 2003 15:10:19 -0400 Received: (from aturner@localhost) by mx1.neteconomist.com (8.12.8/8.12.8/Submit) id h8BJAIT3016437; Thu, 11 Sep 2003 15:10:18 -0400 Date: Thu, 11 Sep 2003 15:10:18 -0400 From: aturner@neteconomist.com To: Tom Lane Cc: Richard Huxton , pgsql-performance@postgresql.org Subject: Re: Upgrade Woes Message-ID: <20030911191018.GA11764@neteconomist.com> References: <20030910175340.GC5506@neteconomist.com> <200309101931.53746.dev@archonet.com> <20030911141622.GB26759@neteconomist.com> <27548.1063295252@sss.pgh.pa.us> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <27548.1063295252@sss.pgh.pa.us> User-Agent: Mutt/1.4.1i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200309/193 X-Sequence-Number: 3559 In the performance case the machine was running RedHat AS 2.1. I have posted the database schema at (obtained from pg_dump -s): http://serverbeach.plexq.com/~aturner/schema.sql The time to run all the stats procedures dropped through the floor. refresh_hourly_iud, adl_hourly_iud, rebuild_daily_total etc. There is a python script that calls the proc once for each hour or day. When running the historical calc job for a 7 day period back, it would crawl on 7.3.3. We started benching the drive array and found other issues with the system in the mean time (like the drive array was giving us 10MB/sec write speed - the guy who set it up did not enable write to cache). Once it was reconfigured the DB performance did not improve much (bonnie++ was used to verify the RAID array speed). Alex Turner On Thu, Sep 11, 2003 at 11:47:32AM -0400, Tom Lane wrote: > aturner@neteconomist.com writes: > > As for 7.3.3, the project in question suffered a 10x performance > > degredation on 7.3.3 which went away when we rolled back to 7.3.2. > > I would like to pursue that report and find out why. I've just gone > through the CVS logs between 7.3.2 and 7.3.3, and I don't see any change > that would explain a 10x slowdown. Can you provide more details about > exactly what slowed down? > > Also, what PG version were you using on the old RedHat 7.2 installation? > > regards, tom lane