Message-ID: <48291BAF.1040807@commandprompt.com>
Date: Mon, 12 May 2008 21:40:15 -0700
From: "Joshua D. Drake" <jd@commandprompt.com>
Organization: Command Prompt, Inc.
User-Agent: Thunderbird 2.0.0.14 (Windows/20080421)
MIME-Version: 1.0
To: Chris Ruprecht <chris@ruprecht.org>
CC: Francisco Reyes <lists@stringsutils.com>,
	Pgsql performance <pgsql-performance@postgresql.org>
Subject: Re: RAID controllers for Postgresql on large setups
References: <cone.1210644243.416947.84940.1000@zoraida.natserv.net>
	<20080512191118.1b40324b@jd-laptop>
	<87F7C977-3160-42BC-98CF-5271B972C0D9@ruprecht.org>
In-Reply-To: <87F7C977-3160-42BC-98CF-5271B972C0D9@ruprecht.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Chris Ruprecht wrote:
> Joshua,
> 
> did you try to run the 345 on an IBM ServeRAID 6i?

No the only controllers I had at the time were the 2120 and the LSI on 
board that is limited to RAID 1. I put the drives on the LSI in JBOD and 
used Linux software raid.

The key identifier for me was using a single writer over 6 (RAID 10) 
drives with the 2120 I could get ~ 16 megs a second. The moment I went 
to multiple writers it dropped exponentially.

However with software raid I was able to sustain ~ 16 megs a second over 
multiple threads. I stopped testing at 4 threads when I was getting 16 
megs per thread :). I was happy at that point.


Joshua D. Drake