Received: from localhost (unknown [200.46.204.183]) by postgresql.org (Postfix) with ESMTP id D584965005C for ; Thu, 21 Aug 2008 16:15:25 -0300 (ADT) Received: from postgresql.org ([200.46.204.86]) by localhost (mx1.hub.org [200.46.204.183]) (amavisd-maia, port 10024) with ESMTP id 09269-03 for ; Thu, 21 Aug 2008 16:14:57 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey-1.7.6 Received: from smtp101.biz.mail.mud.yahoo.com (smtp101.biz.mail.mud.yahoo.com [68.142.200.236]) by postgresql.org (Postfix) with SMTP id 5FB8364FF7E for ; Thu, 21 Aug 2008 16:14:51 -0300 (ADT) Received: (qmail 2527 invoked from network); 21 Aug 2008 19:14:49 -0000 Received: from unknown (HELO ?192.168.5.112?) (ramayer@forensiclogic.com@76.201.143.184 with plain) by smtp101.biz.mail.mud.yahoo.com with SMTP; 21 Aug 2008 19:14:49 -0000 X-YMail-OSG: DN0ZREAVM1n0ib_L57rULmM01w9JkQvFk35k9U8Q2_ygB_obnt5S9uv8OEda8hVvdbR6343HohHBs0O06tQey5ykCQgFKitNsKq5l79gYUOK3o0SyfM9TDRhsxflybLSzds- X-Yahoo-Newman-Property: ymail-3 Message-ID: <48ADBEA7.5060107@cheapcomplexdevices.com> Date: Thu, 21 Aug 2008 12:14:47 -0700 From: Ron Mayer User-Agent: Thunderbird 2.0.0.16 (X11/20080724) MIME-Version: 1.0 To: Scott Carey CC: Christiaan Willemsen , "Joshua D. Drake" , Merlin Moncure , pgsql-performance@postgresql.org Subject: Re: How to setup disk spindles for best performance References: <972BDF1A-B4ED-46EC-B416-0CDF15A0439A@technocon.com> <331B8FD1-7460-4B97-BD2C-03D8381E1C4D@technocon.com> <48AD022E.7090400@commandprompt.com> <48AD288B.8030902@technocon.com> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: Maia Mailguard 1.0.1 X-Spam-Status: No, hits=0 tagged_above=0 required=5 tests=none X-Spam-Level: X-Archive-Number: 200808/258 X-Sequence-Number: 31139 Scott Carey wrote: > For reads, if your shared_buffers is large enough, your heavily used > indexes won't likely go to disk much at all. ISTM this would happen regardless of your shared_buffers setting. If you have enough memory the OS should cache the frequently used pages regardless of shared_buffers; and if you don't have enough memory it won't. > ... splitting up data and indexes ... FWIW, I've had a system where moving pgsql_tmp to different disks helped more than moving indexes.