Received: from localhost (maia-3.hub.org [200.46.204.184]) by postgresql.org (Postfix) with ESMTP id CCD299FB2AC for ; Thu, 7 Jun 2007 06:27:59 -0300 (ADT) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.184]) (amavisd-maia, port 10024) with ESMTP id 15198-07 for ; Thu, 7 Jun 2007 06:27:47 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey-1.7.5 Received: from mail.metronet.co.uk (mail.metronet.co.uk [213.162.97.75]) by postgresql.org (Postfix) with ESMTP id 8846A9FA9BD for ; Thu, 7 Jun 2007 06:27:52 -0300 (ADT) Received: from server3.office.archonet.com (84-51-143-99.archon037.adsl.metronet.co.uk [84.51.143.99]) by smtp.metronet.co.uk (MetroNet Mail) with ESMTP id E75181438C9; Thu, 7 Jun 2007 10:23:19 +0100 (BST) Received: from [192.168.1.36] (dell36.office.archonet.com [192.168.1.36]) by server3.office.archonet.com (Postfix) with ESMTP id 9BD6B274154; Thu, 7 Jun 2007 10:23:19 +0100 (BST) Message-ID: <4667CE87.30101@archonet.com> Date: Thu, 07 Jun 2007 10:23:19 +0100 From: Richard Huxton User-Agent: Thunderbird 2.0.0.0 (X11/20070326) MIME-Version: 1.0 To: Mark Kirkwood Cc: Kurt Overberg , pgsql-performance@postgresql.org Subject: Re: Weird 8.2.4 performance References: <371D1936-E96B-4865-AEEF-C82AC41C2982@hotdogrecords.com> <46674AEF.6040803@paradise.net.nz> In-Reply-To: <46674AEF.6040803@paradise.net.nz> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: Maia Mailguard 1.0.1 X-Archive-Number: 200706/120 X-Sequence-Number: 25153 Mark Kirkwood wrote: > > 8.2 is deciding to use a bitmap index scan on taskid_taskinstance_key, > which seems to be slower (!) than a plain old index scan that 8.0 is > using. A dirty work around is to disable bitmap scans via: I'm having difficulty figuring out why it's doing this at all. There's only one index involved, and it's over the primary-key to boot! An EXPLAIN ANALYSE with enable_bitmapscan off should say why PG thinks the costs are cheaper than they actually are. PS - well worded question Kurt. All the relevant information neatly laid out, explain analyse on both platforms - you should be charging to let people help ;-) -- Richard Huxton Archonet Ltd