X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (mx2.hub.org [200.46.204.254]) by postgresql.org (Postfix) with ESMTP id 1AA0D11F65D7 for ; Fri, 14 Apr 2006 17:09:37 -0300 (ADT) X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Score: -0.11 X-Spam-Level: X-Spam-Status: No, score=-0.11 required=5 tests=[AWL=-1.462, BAYES_00=-2.599, FORGED_RCVD_HELO=0.135, SPF_HELO_SOFTFAIL=2.432, SPF_SOFTFAIL=1.384] Received: from postgresql.org ([200.46.204.71]) by localhost (mx2.hub.org [200.46.204.254]) (amavisd-new, port 10024) with ESMTP id TbjWwOBVkwbF for ; Fri, 14 Apr 2006 17:09:20 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey- Received: from zoraida.natserv.net (p65-147.acedsl.com [66.114.65.147]) by postgresql.org (Postfix) with ESMTP id 64B9711F65C0 for ; Fri, 14 Apr 2006 17:09:19 -0300 (ADT) Received: from zoraida.natserv.net (localhost.natserv.net [127.0.0.1]) by zoraida.natserv.net (Postfix) with ESMTP id 45BABB823; Fri, 14 Apr 2006 16:09:19 -0400 (EDT) Received: from 35st-server.simplicato.com (static-71-249-233-130.nycmny.east.verizon.net [71.249.233.130]) by zoraida.natserv.net (Postfix) with ESMTP id E89B3B81F; Fri, 14 Apr 2006 16:09:18 -0400 (EDT) References: <443DA715.4060606@gmail.com> <9549.1144903344@sss.pgh.pa.us> <20060414055923.GX49405@pervasive.com> <20060414114911.GG32678@mathom.us> <20060414190620.GH32678@mathom.us> Message-ID: X-Mailer: http://www.courier-mta.org/cone/ From: Francisco Reyes To: Michael Stone Cc: pgsql-performance@postgresql.org Subject: Re: Inserts optimization? Date: Fri, 14 Apr 2006 16:09:18 -0400 Mime-Version: 1.0 Content-Type: text/plain; format=flowed; charset="US-ASCII" Content-Disposition: inline Content-Transfer-Encoding: 7bit X-Virus-Scanned: ClamAV using ClamSMTP X-Archive-Number: 200604/344 X-Sequence-Number: 18428 Michael Stone writes: > I still don't follow that. Why would the RAID level matter? IOW, are you > actually wanting 2 spares, or are you just stick with that because you > need a factor of two disks for your mirrors? RAID 10 needs pairs.. so we can either have no spares or 2 spares. > Mmm, it's a bit more complicated than that. RAID 10 can be better if you > have lots of random writes (though a large RAID cache can mitigate > that). We are using a 3ware 9550SX with 128MB RAM (at least I believe that is what that card has installed). >For small random reads the limiting factor is how >fast you can seek, and that number is based more on the number of disks than the RAID > level. I don't have any solid stats, but I would guess the machines will fairly close split between reads and writes. > hardware. The reason that RAID 10 can give better random small block > write performance is that fewer disks need to be involved per write. That makes sense. > That's something that can be mitigated with a large cache 128MB enough in your opinion? > the writes, but some controllers are much better than others in that > regard. The controller we are using is 3Ware 9550SX. > This is really a case where you have to test with your > particular hardware & data That is obviously the ideal way to go, but it is very time consuming. :-( To setup a machine with one set of raid setup.. test, then re-do with different set of raid.. re test.. that's anywhere from 1 to 2 days worth of testing. Unlikely I will be given that time to test.