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 2D47F563C1 for ; Mon, 14 Mar 2005 21:31:10 +0000 (GMT) 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 67322-07 for ; Mon, 14 Mar 2005 21:31:05 +0000 (GMT) Received: from wproxy.gmail.com (wproxy.gmail.com [64.233.184.195]) by svr1.postgresql.org (Postfix) with ESMTP id E9B855547E for ; Mon, 14 Mar 2005 21:31:02 +0000 (GMT) Received: by wproxy.gmail.com with SMTP id 70so1852436wra for ; Mon, 14 Mar 2005 13:31:04 -0800 (PST) DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:references; b=KUawNjjQW5ey8aOAPpy5iLzaUoBdhWM32pHAsMyUpzxy2/yLw3F1EN7nSitdN7dkdccKJ09nroeLDpFXIwROvOI0H8ADBTeSNboIQoNtLrpNkWlxRVdrMF1lxglDbHTYKXoDMSqr9wO07qICmdE0vzBS72DY5yw5Tb4iwgJFDHY= Received: by 10.54.7.54 with SMTP id 54mr215573wrg; Mon, 14 Mar 2005 13:31:04 -0800 (PST) Received: by 10.54.18.53 with HTTP; Mon, 14 Mar 2005 13:31:03 -0800 (PST) Message-ID: <33c6269f05031413315458d3cb@mail.gmail.com> Date: Mon, 14 Mar 2005 16:31:03 -0500 From: Alex Turner Reply-To: Alex Turner To: Arshavir Grigorian Subject: Re: Postgres on RAID5 Cc: pgsql-performance@postgresql.org In-Reply-To: <4235FA0A.40600@m-cam.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit References: <423209E1.6090200@m-cam.com> <33c6269f05031119506fee5c31@mail.gmail.com> <4235FA0A.40600@m-cam.com> X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, hits=0.228 tagged_above=0 required=5 tests=AWL, DNS_FROM_RFC_ABUSE, RCVD_BY_IP X-Spam-Level: X-Archive-Number: 200503/286 X-Sequence-Number: 10991 Actualy my statistics were off a bit I realised - chance of failure for one drive is 1 in X. change of failure in RAID 0 is 7 in X, chance of one drive failure in 14 drive RAID 5 is 14 in X,13 in X for second drive, total probably is 182 in X*X, which is much lower than RAID 0. Your drive performance is less than stellar for a 14 drive stripe, and CPU usage for writes is very high. Even so - this should be enough through put to get over 100 rows/sec assuming you have virtualy no stored procs (I have noticed that stored procs in plpgsql REALLY slow pg_sql down). Alex Turner netEconomist On Mon, 14 Mar 2005 15:54:34 -0500, Arshavir Grigorian wrote: > Alex Turner wrote: > > I would recommend running a bonnie++ benchmark on your array to see if > > it's the array/controller/raid being crap, or wether it's postgres. I > > have had some very surprising results from arrays that theoretically > > should be fast, but turned out to be very slow. > > > > I would also seriously have to recommend against a 14 drive RAID 5! > > This is statisticaly as likely to fail as a 7 drive RAID 0 (not > > counting the spare, but rebuiling a spare is very hard on existing > > drives). > > Thanks for the reply. > > Here are the results of the bonnie test on my array: > > ./bonnie -s 10000 -d . > oo 2>&1 > File './Bonnie.23736', size: 10485760000 > Writing with putc()...done > Rewriting...done > Writing intelligently...done > Reading with getc()...done > Reading intelligently...done > Seeker 1...Seeker 2...Seeker 3...start 'em...done...done...done... > -------Sequential Output-------- ---Sequential Input-- --Random-- > -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks--- > MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU > 10000 4762 96.0 46140 78.8 31180 61.0 3810 99.9 71586 67.7 411.8 13.1 > > On a different note, I am not sure how the probability of RAID5 over 15 > disks failing is the same as that of a RAID0 array over 7 disks. RAID5 > can operate in a degraded mode (14 disks - 1 bad), RAID0 on the other > hand cannot operate on 6 disks (6 disks - 1 bad). Am I missing something? > > Are you saying running RAID0 on a set of 2 RAID1 arrays of 7 each? That > would work fine, except I cannot afford to "loose" that much space. > > Care to comment on these numbers? Thanks. > > > Arshavir >