Received: from localhost (unknown [200.46.204.184]) by postgresql.org (Postfix) with ESMTP id 1D5322E006B for ; Mon, 3 Mar 2008 15:45:10 -0400 (AST) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.184]) (amavisd-maia, port 10024) with ESMTP id 87189-05 for ; Mon, 3 Mar 2008 15:45:01 -0400 (AST) X-Greylist: delayed 00:06:39.548495 by SQLgrey-1.7.5 Received: from smtp110.biz.mail.mud.yahoo.com (smtp110.biz.mail.mud.yahoo.com [68.142.201.179]) by postgresql.org (Postfix) with SMTP id DECBB2E0056 for ; Mon, 3 Mar 2008 15:45:03 -0400 (AST) Received: (qmail 71009 invoked from network); 3 Mar 2008 19:38:22 -0000 Received: from unknown (HELO ?192.168.5.105?) (ramayer@forensiclogic.com@75.37.20.136 with plain) by smtp110.biz.mail.mud.yahoo.com with SMTP; 3 Mar 2008 19:38:22 -0000 X-YMail-OSG: 1rCm118VM1ltIyH6cSL.AyK6nsTQbfDRpBQCTj43n0rrN10GylbUxUS91BqbMmg35K9kzgefHt1rXtGCebQIavveXDxPY1hEkuJRv6LJKISHXKa1l78cHMBpEEGrrg-- X-Yahoo-Newman-Property: ymail-3 Message-ID: <47CC53AD.6090602@cheapcomplexdevices.com> Date: Mon, 03 Mar 2008 11:38:21 -0800 From: Ron Mayer User-Agent: Mozilla-Thunderbird 2.0.0.9 (X11/20080110) MIME-Version: 1.0 To: Greg Smith CC: Steve Poe , pgsql-performance@postgresql.org Subject: Re: How to choose a disc array for Postgresql? References: <721b21dc0803012337u55ddaa13l5517816659f8e100@mail.gmail.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: 200803/37 X-Sequence-Number: 28991 Greg Smith wrote: > On Sat, 1 Mar 2008, Steve Poe wrote: >> SATA over SCSI? > > I've collected up many of the past list comments on this subject and put > a summary at > http://www.postgresqldocs.org/index.php/SCSI_vs._IDE/SATA_Disks Should this section: ATA Disks... Always default to the write cache enabled....turn it off.... be amended to say that if you have an OS that supports write barriers (linuxes newer than early 2005) you shouldn't worry about this? And perhaps the SCSI section should also be amended to say that that the same 2.6 kernels that fail to send the IDE FLUSH CACHE command also fail to send the SCSI SYNCHRONIZE CACHE command, so you should go through the same cache-disabling hoops there? References from the Linux SATA driver guy and lwn here: http://hardware.slashdot.org/comments.pl?sid=149349&cid=12519114 http://lwn.net/Articles/77074/