X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (av.hub.org [200.46.204.144]) by svr1.postgresql.org (Postfix) with ESMTP id A78EDDBAB9 for ; Fri, 18 Nov 2005 13:31:59 -0400 (AST) 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 63310-02 for ; Fri, 18 Nov 2005 17:31:58 +0000 (GMT) X-Greylist: from auto-whitelisted by SQLgrey- Received: from unicorn.rentec.com (unicorn.rentec.com [216.223.240.9]) by svr1.postgresql.org (Postfix) with ESMTP id F1906DBA9A for ; Fri, 18 Nov 2005 13:31:56 -0400 (AST) Received: from ram.rentec.com (mailhost [192.5.35.66]) by unicorn.rentec.com (8.13.1/8.12.1) with ESMTP id jAIHVTNL027900 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Fri, 18 Nov 2005 12:31:31 -0500 (EST) X-Source: non-mednet Received: from [172.26.132.145] (hoopoe.rentec.com [172.26.132.145]) by ram.rentec.com (8.13.1/8.12.1) with ESMTP id jAIHUwrB012673; Fri, 18 Nov 2005 12:30:58 -0500 (EST) Message-ID: <437E0FF5.5090203@rentec.com> Date: Fri, 18 Nov 2005 12:31:33 -0500 From: Alan Stange Reply-To: stange@rentec.com Organization: Renaissance Technologies Corp. User-Agent: Thunderbird 1.5 (X11/20051025) MIME-Version: 1.0 To: Luke Lonergan CC: Dave Cramer , Greg Stark , Joshua Marsh , pgsql-performance@postgresql.org Subject: Re: Hardware/OS recommendations for large databases ( References: In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Logged: Logged by unicorn.rentec.com as jAIHVTNL027900 at Fri Nov 18 12:31:31 2005 X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, score=0.027 required=5 tests=[AWL=0.027] X-Spam-Score: 0.027 X-Spam-Level: X-Archive-Number: 200511/361 X-Sequence-Number: 15618 Luke Lonergan wrote: > Alan, > > On 11/18/05 8:13 AM, "Alan Stange" wrote: > > I told you in my initial post that I was observing numbers in > excess of > what you claiming, but you seemed to think I didn't know how to > measure > an IO rate. > > Prove me wrong, post your data. > > I should note too that our system uses about 20% of a single cpu when > performing a table scan at >100MB/s of IO. I think you claimed the > system would be cpu bound at this low IO rate. > > > See above. Here's the output from one iteration of iostat -k 60 while the box is doing a select count(1) on a 238GB table. avg-cpu: %user %nice %sys %iowait %idle 0.99 0.00 17.97 32.40 48.64 Device: tps kB_read/s kB_wrtn/s kB_read kB_wrtn sdd 345.95 130732.53 0.00 7843952 0 We're reading 130MB/s for a full minute. About 20% of a single cpu was being used. The remainder being idle. We've done nothing fancy and achieved results you claim shouldn't be possible. This is a system that was re-installed yesterday, no tuning was done to the file systems, kernel or storage array. What am I doing wrong? 9 years ago I co-designed a petabyte data store with a goal of 1GB/s IO (for a DOE lab). And now I don't know what I'm doing, Cheers, -- Alan