X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.144]) by svr1.postgresql.org (Postfix) with ESMTP id 5C47ED1B1FD for ; Tue, 20 Jul 2004 12:27:49 -0300 (ADT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (av.hub.org [200.46.204.144]) (amavisd-new, port 10024) with ESMTP id 56815-08 for ; Tue, 20 Jul 2004 15:27:46 +0000 (GMT) Received: from mpls-qmqp-03.inet.qwest.net (mpls-qmqp-03.inet.qwest.net [63.231.195.114]) by svr1.postgresql.org (Postfix) with SMTP id B011AD1B1E0 for ; Tue, 20 Jul 2004 12:27:44 -0300 (ADT) Received: (qmail 96583 invoked by uid 0); 20 Jul 2004 14:57:08 -0000 Received: from mpls-pop-03.inet.qwest.net (63.231.195.3) by mpls-qmqp-03.inet.qwest.net with QMQP; 20 Jul 2004 14:57:08 -0000 Received: from 63-227-127-37.dnvr.qwest.net (HELO ?10.0.0.2?) (63.227.127.37) by mpls-pop-03.inet.qwest.net with SMTP; 20 Jul 2004 15:27:43 -0000 Date: Tue, 20 Jul 2004 09:28:54 -0600 Message-Id: <1090337334.709.21.camel@localhost.localdomain> From: "Scott Marlowe" To: bsimon@loxane.com Cc: pgsql-performance@postgresql.org Subject: Re: =?ISO-8859-1?Q?R=E9f=2E?= : Re: NAS, SAN or any In-Reply-To: References: Content-Type: text/plain Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.6 (1.4.6-2) Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, hits=0.1 tagged_above=0.0 required=5.0 tests=RCVD_IN_SORBS X-Spam-Level: X-Archive-Number: 200407/127 X-Sequence-Number: 7518 On Tue, 2004-07-20 at 03:32, bsimon@loxane.com wrote: > Thanks a lot Scott. > > It seems that we were totally wrong when considering a network storage > solution. I've read your techdoc > http://techdocs.postgresql.org/guides/DiskTuningGuide and found many > interesting remarks. > I think that we will know focus on external Raid systems which seem to > be relativily affordable compared to NAS or SAN (we would have had the > budget for one of these). > As we don't plan to have more than 5 connections (I.E process), we > think SATA drives would fit our requirements. Could this be an issue > for an after crash recovery ? If you're looking at (S)ATA RAID, definitely look at escalade, as another poster mentioned. Last year I and a few other folks on the lists were testing RAID controllers for survival of the power plug pull test, and the Escalade passed (someone else did the testing, I tested the LSI MegaRAID 320-2 controller with battery backed cache). > We also hesitate concerning the raid level to use. We are currently > comparing raid 1+0 and raid 5 but we have no actual idea on which one > to use. > > Our priorities are : > 1) performance > 2) recovery > 3) price > 4) back-up Basically, for a smaller number of drivers, RAID 1+0 is almost always a win over RAID 5. As the number of drives in the array grows, RAID 5 usually starts to pull back in the lead. RAID 5 definitely gives you the most storage for your dollar of any of the redundant array types. The more important point of a RAID controller is that it have battery backed cache to make sure that the database server isn't waiting for WAL writes all the time. A single port LSI Megaraid 320-1 controller is only about $500 or less, the last time I checked (with battery backed cache, order it WITH the battery and cache, otherwise you may have a hard time finding the right parts later on.) It supports hot spares for automatic rebuild.