From owner-freebsd-performance@FreeBSD.ORG Sat Jan 13 11:53:50 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id C3EBD16A403 for ; Sat, 13 Jan 2007 11:53:50 +0000 (UTC) (envelope-from michel@lucenet.com.br) Received: from msrv.matik.com.br (msrv.matik.com.br [200.152.83.14]) by mx1.freebsd.org (Postfix) with ESMTP id 3453513C442 for ; Sat, 13 Jan 2007 11:53:49 +0000 (UTC) (envelope-from michel@lucenet.com.br) Received: from webmail.matik.com.br (localhost [127.0.0.1]) by msrv.matik.com.br (8.13.8/8.13.1) with ESMTP id l0DBrkOJ017508; Sat, 13 Jan 2007 09:53:47 -0200 (BRST) (envelope-from michel@lucenet.com.br) Received: from 200.152.83.36 (SquirrelMail authenticated user luc.michel) by webmail.matik.com.br with HTTP; Sat, 13 Jan 2007 09:53:47 -0200 (BRST) Message-ID: <63758.200.152.83.36.1168689227.squirrel@webmail.matik.com.br> In-Reply-To: <45A87878.1050505@paradise.net.nz> References: <64656.200.152.83.36.1168651673.squirrel@webmail.matik.com.br> <45A87878.1050505@paradise.net.nz> Date: Sat, 13 Jan 2007 09:53:47 -0200 (BRST) From: "Michel Santos" To: "Mark Kirkwood" User-Agent: SquirrelMail/1.4.9a MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Priority: 3 (Normal) Importance: Normal Cc: freebsd-performance@freebsd.org Subject: Re: diskio low read performance X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 13 Jan 2007 11:53:50 -0000 Mark Kirkwood disse na ultima mensagem: > Michel Santos wrote: >> I hope somebody can help me. I upgraded recently my cache server (squid) >> 4.11 to 6.2 >> >> I notice that my disk read performance seems to be low. I observe it by >> reading diskio from net-snmp. I started to check this because my old >> server gave much better results from cache. >> >> I see read access as 1/4 - 1/6 of write access. My cache_dirs are full >> and >> no swap in use. IOLA is not over 25-30 in peaks >> >> is there some special configuration to get better read performance on >> 6.2? >> My disks are UW320 10k on Adaptec 29320 and the same as on the 4.11 >> server. I really dont care about write speed so much so if there are >> parameters to configure I apreciate to hear about. >> >> I checked the following with 1 and 0 but it does not make any difference >> at all >> >> vfs.write_behind >> vfs.vmiodirenable >> >> I get an average of 6-8Mb/s through this server and 3-4 of it is http >> traffic. I see the object are going into the cache but for some reason >> it >> seems they can not be read in time. >> >> I tried different newfs -b and -f and actually 1024/4096 seems to be >> best. >> I do not cache large files on this machine. >> >> >> Is there something I sgould try? >> > > Note sure this will help you, as your files are all small, but try > increasing vfs.read_max (say 16 or 32). > I forgot to say that I tried it already. Even if it gave me no improvement I have it in 16 at this time together with a higher vfs.ufs.dirhash_maxmem value. Sincerley, any of the configuration changes I did gave me absolutely nothing in relationship to the disk read access performance. That is disappointing. Should I go back and try ufs1 perhaps? Or is it that squid does not work well on 6.2? I use diskd but in order to check if there is a SHM issue I tried ufs and aufs, but also no difference at all. On 6.2 I do not even get close to 50% of 4.11 disk read performance. thank you Michel computador é como nem cavalo e mulher mais que montam neles, pior que ficam ... **************************************************** Datacenter Matik http://datacenter.matik.com.br E-Mail e Data Hosting Service para Profissionais. ****************************************************