X-Original-To: pgsql-performance-postgresql.org@postgresql.org Received: from localhost (mx1.hub.org [200.46.208.251]) by postgresql.org (Postfix) with ESMTP id 285159FA383 for ; Wed, 13 Sep 2006 13:08:00 -0300 (ADT) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.208.251]) (amavisd-new, port 10024) with ESMTP id 53028-07 for ; Wed, 13 Sep 2006 13:07:55 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey- Received: from servata.com (ip-216-152-249-241.servata.com [216.152.249.241]) by postgresql.org (Postfix) with ESMTP id BDD509FA374 for ; Wed, 13 Sep 2006 13:07:55 -0300 (ADT) Received: from [209.162.219.253] (helo=dogma.v10.wvs) by servata.com with esmtpa (Exim 4.50) id 1GNXGz-0002ca-8q; Wed, 13 Sep 2006 09:07:37 -0700 Subject: Re: Performance With Joins on Large Tables From: Jeff Davis To: Joshua Marsh Cc: pgsql-performance@postgresql.org, jim@nasby.net In-Reply-To: <38242de90609130749u976295fwb10904ff4797b920@mail.gmail.com> References: <38242de90609121517y62517040v8622b9ae93cba23d@mail.gmail.com> <20060913064152.GC23173@nasby.net> <38242de90609130749u976295fwb10904ff4797b920@mail.gmail.com> Content-Type: text/plain Date: Wed, 13 Sep 2006 09:07:34 -0700 Message-Id: <1158163655.29889.57.camel@dogma.v10.wvs> Mime-Version: 1.0 X-Mailer: Evolution 2.0.4 (2.0.4-7) Content-Transfer-Encoding: 7bit X-Virus-Scanned: Maia Mailguard 1.0.1 X-Spam-Status: No, hits=0.238 tagged_above=0 required=5 tests=AWL, FORGED_RCVD_HELO, SPF_HELO_PASS X-Spam-Level: X-Archive-Number: 200609/121 X-Sequence-Number: 20807 On Wed, 2006-09-13 at 08:49 -0600, Joshua Marsh wrote: > That seems to have done it. Are there any side effects to this > change? I read about random_page_cost in the documentation and it > seems like this is strictly for planning. All the tables on this > database will be indexed and of a size similar to these two, so I > don't see it causing any other problems. Though I would check though > :) > Right, it's just used for planning. Avoid setting it too low, if it's below about 2.0 you would most likely see some very strange plans. Certainly it doesn't make sense at all to set it below 1.0, since that is saying it's cheaper to get a random page than a sequential one. What was your original random_page_cost, and what is the new value you set it to? Regards, Jeff Davis