From owner-freebsd-geom@FreeBSD.ORG Sat Dec 30 06:45:00 2006 Return-Path: X-Original-To: freebsd-geom@freebsd.org Delivered-To: freebsd-geom@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 0CBBD16A403 for ; Sat, 30 Dec 2006 06:45:00 +0000 (UTC) (envelope-from anderson@centtech.com) Received: from mh1.centtech.com (moat3.centtech.com [64.129.166.50]) by mx1.freebsd.org (Postfix) with ESMTP id AE57213C428 for ; Sat, 30 Dec 2006 06:44:59 +0000 (UTC) (envelope-from anderson@centtech.com) Received: from [192.168.42.21] (andersonbox1.centtech.com [192.168.42.21]) by mh1.centtech.com (8.13.8/8.13.8) with ESMTP id kBU6UPeZ012983; Sat, 30 Dec 2006 00:30:26 -0600 (CST) (envelope-from anderson@centtech.com) Message-ID: <45960781.3040304@centtech.com> Date: Sat, 30 Dec 2006 00:30:25 -0600 From: Eric Anderson User-Agent: Thunderbird 1.5.0.9 (X11/20061223) MIME-Version: 1.0 To: Ivan Voras References: <20061228171858.GA11296@qlovarnika.bg.datamax> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: ClamAV 0.88.4/2395/Fri Dec 29 17:16:50 2006 on mh1.centtech.com X-Virus-Status: Clean X-Spam-Status: No, score=-2.5 required=8.0 tests=AWL,BAYES_00 autolearn=ham version=3.1.6 X-Spam-Checker-Version: SpamAssassin 3.1.6 (2006-10-03) on mh1.centtech.com Cc: freebsd-geom@freebsd.org Subject: Re: gstripe performance scaling with many disks X-BeenThere: freebsd-geom@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: GEOM-specific discussions and implementations List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 30 Dec 2006 06:45:00 -0000 On 12/28/06 12:27, Ivan Voras wrote: > Ivan Voras wrote: > >> I'll take a shot at this: Since maximum kernel reads are still limited >> to 128 KB/s, by adding more drives you're making individual requests >> shorter. I.e. with one drive, it gets 128 KB requests, with two, each >> gets 64 KB, with 16, each gets 8 KB. So network & kernel latency becomes >> visible. > > Btw. if I'm right, you should be able to observe this with gstat and/or > iostat. If you don't see it, then it's possible I'm wrong :) Maybe this should be tried with md* devices just to remove the network component from this mix. Eric -- ------------------------------------------------------------------------ Eric Anderson Sr. Systems Administrator Centaur Technology An undefined problem has an infinite number of solutions. ------------------------------------------------------------------------