Received: from localhost (unknown [200.46.204.184]) by postgresql.org (Postfix) with ESMTP id 4D2532E007E for ; Wed, 19 Mar 2008 10:41:14 -0300 (ADT) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.184]) (amavisd-maia, port 10024) with ESMTP id 39159-04 for ; Wed, 19 Mar 2008 10:41:02 -0300 (ADT) Received: from mx1.hub.org (mx1.hub.org [200.46.204.176]) by postgresql.org (Postfix) with ESMTP id 459AE2E0037 for ; Wed, 19 Mar 2008 10:41:11 -0300 (ADT) X-Greylist: delayed 04:39:35.735302 by SQLgrey-1.7.5 Received: from smtprelay.hostedemail.com (smtprelay0129.hostedemail.com [216.40.44.129]) by mx1.hub.org (Postfix) with ESMTP id 8311916D7CB2 for ; Wed, 19 Mar 2008 10:41:11 -0300 (ADT) Received: from smtprelay.hostedemail.com (ff-bigip1 [10.5.19.254]) by smtpgrave02.hostedemail.com (Postfix) with ESMTP id 3BB1EFA930 for ; Tue, 18 Mar 2008 18:52:22 +0000 (UTC) Received: from emd2-omf02.hostedemail.com (ff-bigip1 [10.5.19.254]) by smtprelay03.hostedemail.com (Postfix) with ESMTP id 007A52582C; Tue, 18 Mar 2008 18:52:17 +0000 (UTC) X-SpamScore: 2 X-Spamcatcher-Summary: 2, 0, 0, 2e0451ddacc4266d, 41e1833f647c59c0, greg@tcscs.com, -, RULES_HIT:152:355:379:476:541:599:601:945:960:962:967:973:983:988:989:1189:1208:1212:1221:1260:1261:1313:1314:1345:1358:1359:1373:1431:1436:1437:1515:1516:1517:1535:1541:1575:1588:1589:1592:1594:1676:1711:1730:1776:1792:2068:2069:2194:2198:2199:2200:2525:2553:2559:2564:2682:2685:2857:2859:2899:2933:2937:2939:2942:2945:2947:2951:2954:3022:3027:3280:3352:3622:3865:3866:3867:3868:3869:3870:3871:3872:3873: 3874:3934:3936:3938:3941:3944:4250:4321:5007:6117:6119:7652:7679:7903, 0, RBL:none, CacheIP:none, Bayesian:0.5, 0.5, 0.5, Netcheck:none, DomainCache:0, MSF:not bulk, SPF:, MSBL:none, DNSBL:none X-Spamcatcher-Explanation: Received: from [10.7.1.42] (unknown [71.39.164.77]) (Authenticated sender: greg@tcscs.com) by emd2-omf02.hostedemail.com (Postfix) with ESMTP; Tue, 18 Mar 2008 18:52:17 +0000 (UTC) Subject: Re: What is the best way to storage music files in Postgresql From: Gregory Youngblood To: david@lang.hm Cc: Andrej Ricnik-Bay , Craig Ringer , pgsql-performance@postgresql.org In-Reply-To: References: <8fb205ea0803152311n2c950758g253f4cefd20cff7d@mail.gmail.com> <47DCBD63.9050109@postnewspapers.com.au> <4544e0330803171101g680ffa5bj8ac32908a9f3dfb5@mail.gmail.com> <8fb205ea0803171139h63b55fdrf62776851583a74d@mail.gmail.com> <4544e0330803172026h3eae59a1u8914bda1a5a95bb5@mail.gmail.com> <47DF9278.70605@postnewspapers.com.au> Content-Type: multipart/alternative; boundary="=-ki1uWBvtw6xoR1DrFgqW" Organization: The Complete Solution Date: Tue, 18 Mar 2008 11:52:16 -0700 Message-Id: <1205866336.483.5.camel@hammy> Mime-Version: 1.0 X-Mailer: Evolution 2.12.1 X-Virus-Scanned: Maia Mailguard 1.0.1 X-Spam-Status: No, hits=2.126 tagged_above=0 required=5 tests=DNS_FROM_RFC_BOGUSMX=2.125, HTML_MESSAGE=0.001 X-Spam-Level: ** X-Archive-Number: 200803/398 X-Sequence-Number: 29352 --=-ki1uWBvtw6xoR1DrFgqW Content-Type: text/plain Content-Transfer-Encoding: 7bit On Tue, 2008-03-18 at 11:04 -0700, david@lang.hm wrote: > On Wed, 19 Mar 2008, Andrej Ricnik-Bay wrote: > > > > > On 18/03/2008, Craig Ringer wrote: > >> Isn't a 10 or 24 spindle RAID 5 array awfully likely to encounter a > >> double disk failure (such as during the load imposed by rebuild onto a > >> spare) ? > > that's why you should use raid6 (allowing for dual failures) And, if possible, try to use drives from a couple of different batches. I wouldn't go crazy trying to do this, but sometimes a combination of drives from two or three batches can make a difference (though not always). Also, a very informative read: http://research.google.com/archive/disk_failures.pdf In short, best thing to do is watch SMART and be prepared to try and swap a drive out before it fails completely. :) --=-ki1uWBvtw6xoR1DrFgqW Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: 7bit On Tue, 2008-03-18 at 11:04 -0700, david@lang.hm wrote: 
On Wed, 19 Mar 2008, Andrej Ricnik-Bay wrote:

> 
> On 18/03/2008, Craig Ringer <craig@postnewspapers.com.au> wrote:
>> Isn't a 10 or 24 spindle RAID 5 array awfully likely to encounter a
>>  double disk failure (such as during the load imposed by rebuild onto a
>>  spare) ?

that's why you should use raid6 (allowing for dual failures)

And, if possible, try to use drives from a couple of different batches. I wouldn't go crazy trying to do this, but sometimes a combination of drives from two or three batches can make a difference (though not always).

Also, a very informative read: http://research.google.com/archive/disk_failures.pdf
In short, best thing to do is watch SMART and be prepared to try and swap a drive out before it fails completely. :) --=-ki1uWBvtw6xoR1DrFgqW--