From owner-freebsd-performance@FreeBSD.ORG Mon Jul 14 06:35:42 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 8BF3F37B40A; Mon, 14 Jul 2003 06:35:42 -0700 (PDT) Received: from mail.drunkencomputing.de (ratz.drunkencomputing.de [195.244.235.248]) by mx1.FreeBSD.org (Postfix) with ESMTP id 3EE6243F85; Mon, 14 Jul 2003 06:34:18 -0700 (PDT) (envelope-from hscholz@raisdorf.net) Received: from localhost (localhost [127.0.0.1]) by mail.drunkencomputing.de (Postfix) with ESMTP id F2E8C8AF6A; Mon, 14 Jul 2003 15:35:24 +0200 (CEST) Received: from mail.drunkencomputing.de ([127.0.0.1])port 10024) with ESMTP id 80356-10; Mon, 14 Jul 2003 15:35:24 +0200 (CEST) Received: from goanna.lan.raisdorf.net (pD9E3CDEE.dip.t-dialin.net [217.227.205.238]) by mail.drunkencomputing.de (Postfix) with ESMTP id 959178AF68; Mon, 14 Jul 2003 15:35:09 +0200 (CEST) Date: Mon, 14 Jul 2003 15:33:50 +0200 From: Hendrik Scholz To: freebsd-doc@freebsd.org Message-Id: <20030714153350.52a0c952.hscholz@raisdorf.net> Followup-To: freebsd-doc@freebsd.org X-Mailer: Sylpheed version 0.8.10claws (GTK+ 1.2.10; i386-portbld-freebsd5.1) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at drunkencomputing.de cc: freebsd-performance@freebsd.org Subject: Network performance tuning article X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 14 Jul 2003 13:35:43 -0000 Hi! I've started a document about performance tuning some time ago and as of now it is in a state where others could benefit from it/contribute to it. I've put all documents on http://home.raisdorf.net/public/networkperf/ http://home.raisdorf.net/public/networkperf/article.html contains the complete article rendered as (ah, you already guessed that :)) HTML. Since this is my first FreeBSD article there might be some formatting errors beside the comments in the source where more work has to be done. Feel free to use it/contribute to it/send patches or flames, ... Hendrik P.S. Follow-Up to freebsd-doc@ set. -- Hendrik Scholz - - http://raisdorf.net/ Forcast for tonight: Dark. From owner-freebsd-performance@FreeBSD.ORG Mon Jul 14 08:05:40 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 7FAE537B404 for ; Mon, 14 Jul 2003 08:05:40 -0700 (PDT) Received: from svaha.com (svaha.com [64.46.156.67]) by mx1.FreeBSD.org (Postfix) with ESMTP id 5EEFB43FAF for ; Mon, 14 Jul 2003 08:05:27 -0700 (PDT) (envelope-from meconlen@obfuscated.net) Received: from obfuscated.net ([64.156.25.5]) (AUTH: LOGIN meconlen, TLS: TLSv1/SSLv3,128bits,RC4-MD5) by svaha.com with esmtp; Mon, 14 Jul 2003 11:05:25 -0400 Message-ID: <3F12C6A0.8060004@obfuscated.net> Date: Mon, 14 Jul 2003 11:05:04 -0400 From: Michael Conlen User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.0.2) Gecko/20030208 Netscape/7.02 X-Accept-Language: en-us, en MIME-Version: 1.0 To: freebsd-performance@freebsd.org Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Subject: Measuring memory perf X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 14 Jul 2003 15:05:40 -0000 I'm running a number of Apache webservers that use 4 GB of memory. I'm working on tuning the configurations for Apache, a number of which affect memory usage. My question is, is there a way to measure inactive pages that get used again. I figure I'll adjust the Apache memory usage while monitoring the inactive use versus the disk usage. -- Michael Conlen From owner-freebsd-performance@FreeBSD.ORG Mon Jul 14 10:39:06 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 891F137B401 for ; Mon, 14 Jul 2003 10:39:06 -0700 (PDT) Received: from out003.verizon.net (out003pub.verizon.net [206.46.170.103]) by mx1.FreeBSD.org (Postfix) with ESMTP id 973E143FBD for ; Mon, 14 Jul 2003 10:39:03 -0700 (PDT) (envelope-from cswiger@mac.com) Received: from mac.com ([141.149.47.46]) by out003.verizon.net (InterMail vM.5.01.05.33 201-253-122-126-133-20030313) with ESMTP id <20030714173902.CPY4805.out003.verizon.net@mac.com>; Mon, 14 Jul 2003 12:39:02 -0500 Message-ID: <3F12EAB0.3050403@mac.com> Date: Mon, 14 Jul 2003 13:38:56 -0400 From: Chuck Swiger Organization: The Courts of Chaos User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.4) Gecko/20030624 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Michael Conlen References: <3F12C6A0.8060004@obfuscated.net> In-Reply-To: <3F12C6A0.8060004@obfuscated.net> X-Enigmail-Version: 0.76.1.0 X-Enigmail-Supports: pgp-inline, pgp-mime Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Authentication-Info: Submitted using SMTP AUTH at out003.verizon.net from [141.149.47.46] at Mon, 14 Jul 2003 12:39:02 -0500 cc: freebsd-performance@freebsd.org Subject: Re: Measuring memory perf X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 14 Jul 2003 17:39:06 -0000 Michael Conlen wrote: > I'm running a number of Apache webservers that use 4 GB of memory. I'm > working on tuning the configurations for Apache, a number of which > affect memory usage. My question is, is there a way to measure inactive > pages that get used again. I figure I'll adjust the Apache memory usage > while monitoring the inactive use versus the disk usage. /usr/ports/web/analog will show you lots of useful stats, including the resources people are hitting most often. Apache is generally stateless and doesn't hold on to resources after serving them, modulo things like mod_perl and the like which do keep state between transactions. -- -Chuck From owner-freebsd-performance@FreeBSD.ORG Mon Jul 14 12:31:44 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 6C0B937B401; Mon, 14 Jul 2003 12:31:44 -0700 (PDT) Received: from otter3.centtech.com (moat3.centtech.com [207.200.51.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 51AA843F3F; Mon, 14 Jul 2003 12:31:43 -0700 (PDT) (envelope-from anderson@centtech.com) Received: from centtech.com (dhcp-243.centtech.com [204.177.173.243]) by otter3.centtech.com (8.12.3/8.12.3) with ESMTP id h6EJVg56041791; Mon, 14 Jul 2003 14:31:42 -0500 (CDT) (envelope-from anderson@centtech.com) Message-ID: <3F13050D.2070100@centtech.com> Date: Mon, 14 Jul 2003 14:31:25 -0500 From: Eric Anderson User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.0.2) Gecko/20030208 Netscape/7.02 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Hendrik Scholz References: <20030714153350.52a0c952.hscholz@raisdorf.net> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-doc@freebsd.org cc: freebsd-performance@freebsd.org Subject: Re: Network performance tuning article X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 14 Jul 2003 19:31:44 -0000 Hendrik Scholz wrote: > Hi! > > I've started a document about performance tuning some time ago and as of > now it is in a state where others could benefit from it/contribute to > it. I've put all documents on > http://home.raisdorf.net/public/networkperf/ > http://home.raisdorf.net/public/networkperf/article.html contains the > complete article rendered as (ah, you already guessed that :)) HTML. > Since this is my first FreeBSD article there might be some formatting > errors beside the comments in the source where more work has to be done. > > Feel free to use it/contribute to it/send patches or flames, ... I had been working on some of this myself, however in a broader sense - covering not just network performance, but any kind of FreeBSD performance tuning. Were you planning on hitting those other areas as well? This could make a big difference on where it should go, or if it should be in a "higher level" section more to itself. Also, I would say that there should be a little more info on "why" things are the way they are - several topics seem to state something, with no reasoning, which in this subject area is very important (I believe). Eric From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 05:51:26 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 6071737B401 for ; Wed, 16 Jul 2003 05:51:26 -0700 (PDT) Received: from spf13.us4.outblaze.com (205-158-62-67.outblaze.com [205.158.62.67]) by mx1.FreeBSD.org (Postfix) with ESMTP id A688343FBD for ; Wed, 16 Jul 2003 05:51:25 -0700 (PDT) (envelope-from jwest254@mail.com) Received: from 205-158-62-68.outblaze.com (205-158-62-68.outblaze.com [205.158.62.68]) by spf13.us4.outblaze.com (Postfix) with QMQP id 3EE371802A2B for ; Wed, 16 Jul 2003 12:51:23 +0000 (GMT) Received: (qmail 94664 invoked from network); 16 Jul 2003 12:51:23 -0000 Received: from unknown (HELO ws1-12.us4.outblaze.com) (205.158.62.81) by 205-158-62-153.outblaze.com with SMTP; 16 Jul 2003 12:51:23 -0000 Received: (qmail 53308 invoked by uid 1001); 16 Jul 2003 12:51:23 -0000 Message-ID: <20030716125123.53307.qmail@mail.com> Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Content-Transfer-Encoding: 7bit MIME-Version: 1.0 X-Mailer: MIME-tools 5.41 (Entity 5.404) Received: from [193.120.103.96] by ws1-12.us4.outblaze.com with http for jwest254@mail.com; Wed, 16 Jul 2003 07:51:22 -0500 From: "James West" To: freebsd-performance@freebsd.org Date: Wed, 16 Jul 2003 07:51:22 -0500 X-Originating-Ip: 193.120.103.96 X-Originating-Server: ws1-12.us4.outblaze.com Subject: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 12:51:26 -0000 Hi, I'm having huge problems with "No buffer space available" errors. I've increased MAXUSERS to 512 in the kernel, recompiled, rebooted and the sysctl values below show that everything is up'ed to the max. some details which may help are: kern.maxfiles: 10796 kern.maxfilesperproc: 9716 kern.ipc.maxsockbuf: 262144 kern.ipc.somaxconn: 65000 kern.ipc.nmbclusters: 65000 kern.ipc.maxsockets: 65000 netstat -m reveals: 70/400/260000 mbufs in use (current/peak/max): 67 mbufs allocated to data 3 mbufs allocated to packet headers 64/208/65000 mbuf clusters in use (current/peak/max) 516 Kbytes allocated to network (0% of mb_map in use) Any executable i try use which uses sockets has problems, for example rsync: socket: No buffer space available rsync: connection unexpectedly closed (0 bytes read so far) rsync error: error in rsync protocol data stream (code 12) at io.c(165) Can anyone help? Regards, James -- __________________________________________________________ Sign-up for your own FREE Personalized E-mail at Mail.com http://www.mail.com/?sr=signup CareerBuilder.com has over 400,000 jobs. Be smarter about your job search http://corp.mail.com/careers From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 05:52:42 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id DA67D37B401 for ; Wed, 16 Jul 2003 05:52:42 -0700 (PDT) Received: from spf13.us4.outblaze.com (205-158-62-67.outblaze.com [205.158.62.67]) by mx1.FreeBSD.org (Postfix) with ESMTP id 5A02743FA3 for ; Wed, 16 Jul 2003 05:52:42 -0700 (PDT) (envelope-from jwest254@mail.com) Received: from 205-158-62-68.outblaze.com (205-158-62-68.outblaze.com [205.158.62.68]) by spf13.us4.outblaze.com (Postfix) with QMQP id 953071810824 for ; Wed, 16 Jul 2003 12:52:40 +0000 (GMT) Received: (qmail 4513 invoked from network); 16 Jul 2003 12:52:39 -0000 Received: from unknown (HELO ws1-12.us4.outblaze.com) (205.158.62.81) by 205-158-62-153.outblaze.com with SMTP; 16 Jul 2003 12:52:39 -0000 Received: (qmail 54509 invoked by uid 1001); 16 Jul 2003 12:52:39 -0000 Message-ID: <20030716125239.54508.qmail@mail.com> Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Content-Transfer-Encoding: 7bit MIME-Version: 1.0 X-Mailer: MIME-tools 5.41 (Entity 5.404) Received: from [193.120.103.96] by ws1-12.us4.outblaze.com with http for jwest254@mail.com; Wed, 16 Jul 2003 07:52:39 -0500 From: "James West" To: freebsd-performance@freebsd.org Date: Wed, 16 Jul 2003 07:52:39 -0500 X-Originating-Ip: 193.120.103.96 X-Originating-Server: ws1-12.us4.outblaze.com Subject: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 12:52:43 -0000 Hi, Sorry for a second mailing - i forgot to mention this is freebsd 4.7-RELEASE and the NIC is a Intel Pro 10/100B/100+ Ethernet Thanks in advance, James. -- __________________________________________________________ Sign-up for your own FREE Personalized E-mail at Mail.com http://www.mail.com/?sr=signup CareerBuilder.com has over 400,000 jobs. Be smarter about your job search http://corp.mail.com/careers From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 06:36:33 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 8512037B404 for ; Wed, 16 Jul 2003 06:36:33 -0700 (PDT) Received: from web14103.mail.yahoo.com (web14103.mail.yahoo.com [216.136.172.133]) by mx1.FreeBSD.org (Postfix) with SMTP id 4CEC643FAF for ; Wed, 16 Jul 2003 06:36:32 -0700 (PDT) (envelope-from cguttesen@yahoo.dk) Message-ID: <20030716133632.32697.qmail@web14103.mail.yahoo.com> Received: from [194.248.174.58] by web14103.mail.yahoo.com via HTTP; Wed, 16 Jul 2003 15:36:32 CEST Date: Wed, 16 Jul 2003 15:36:32 +0200 (CEST) From: =?iso-8859-1?q?Claus=20Guttesen?= To: James West , freebsd-performance@freebsd.org In-Reply-To: <20030716125123.53307.qmail@mail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 13:36:34 -0000 Hi. > I'm having huge problems with "No buffer space > available" errors. I've increased MAXUSERS to 512 in > the kernel, recompiled, rebooted and the sysctl > values below show that everything is up'ed to the > max. > Are you running ipfw on the box? When I configured a queue without a pipe to put it on, I got the same message and couldn't ping hosts on that interface until I removed the queue. regards Claus Yahoo! Mail (http://dk.mail.yahoo.com) - Gratis: 6 MB lagerplads, spamfilter og virusscan From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 06:43:16 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 540DE37B401 for ; Wed, 16 Jul 2003 06:43:16 -0700 (PDT) Received: from venus.veridas.net (venus.veridas.net [203.37.234.51]) by mx1.FreeBSD.org (Postfix) with SMTP id 0670C43FB1 for ; Wed, 16 Jul 2003 06:43:14 -0700 (PDT) (envelope-from lfinch@asitis.net.au) Received: (qmail 19082 invoked by uid 94); 16 Jul 2003 23:37:55 +1000 Received: from lfinch@asitis.net.au by venus by uid 91 with qmail-scanner-1.15 (clamscan: 0.54. Clear:. Processed in 3.879148 secs); 16 Jul 2003 13:37:55 -0000 Received: from nsw-port-203-49-30-225.ausat.net (HELO server2) (203.49.30.225) by ozinet.net.au with SMTP; 16 Jul 2003 23:37:51 +1000 From: "Leigh" To: "'Claus Guttesen'" , "'James West'" , Date: Wed, 16 Jul 2003 23:39:55 +1000 MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: quoted-printable X-Mailer: Microsoft Outlook, Build 11.0.4920 X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 In-Reply-To: <20030716133632.32697.qmail@web14103.mail.yahoo.com> Thread-Index: AcNLn5NRhDDer9kvTAuUyoyQwfZ9TgAACE4w X-Qmail-Scanner-Message-ID: <105836267450219029@venus> Message-Id: <20030716134314.0670C43FB1@mx1.FreeBSD.org> Subject: RE: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 13:43:16 -0000 Another thing you could try is compile for 0 users, this sets it to an automatic value, depending on the machine specs, I had this prob, on a server, and that fixed it. Hope this helps Cheers Leigh -----Original Message----- From: owner-freebsd-performance@freebsd.org [mailto:owner-freebsd-performance@freebsd.org] On Behalf Of Claus = Guttesen Sent: Wednesday, July 16, 2003 11:37 PM To: James West; freebsd-performance@freebsd.org Hi. > I'm having huge problems with "No buffer space > available" errors. I've increased MAXUSERS to 512 in > the kernel, recompiled, rebooted and the sysctl > values below show that everything is up'ed to the > max. > =20 Are you running ipfw on the box? When I configured a queue without a pipe to put it on, I got the same message and couldn't ping hosts on that interface until I removed the queue. regards Claus Yahoo! Mail (http://dk.mail.yahoo.com) - Gratis: 6 MB lagerplads, = spamfilter og virusscan _______________________________________________ freebsd-performance@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-performance To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 07:30:14 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 633D337B401 for ; Wed, 16 Jul 2003 07:30:14 -0700 (PDT) Received: from spf13.us4.outblaze.com (205-158-62-67.outblaze.com [205.158.62.67]) by mx1.FreeBSD.org (Postfix) with ESMTP id 9650043FBD for ; Wed, 16 Jul 2003 07:30:11 -0700 (PDT) (envelope-from jwest254@mail.com) Received: from 205-158-62-68.outblaze.com (205-158-62-68.outblaze.com [205.158.62.68]) by spf13.us4.outblaze.com (Postfix) with QMQP id 847171813B2F for ; Wed, 16 Jul 2003 14:30:11 +0000 (GMT) Received: (qmail 99887 invoked from network); 16 Jul 2003 14:30:10 -0000 Received: from unknown (HELO ws1-6.us4.outblaze.com) (205.158.62.53) by 205-158-62-153.outblaze.com with SMTP; 16 Jul 2003 14:30:10 -0000 Received: (qmail 20980 invoked by uid 1001); 16 Jul 2003 14:30:09 -0000 Message-ID: <20030716143009.20979.qmail@mail.com> Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Content-Transfer-Encoding: 7bit MIME-Version: 1.0 X-Mailer: MIME-tools 5.41 (Entity 5.404) Received: from [193.120.103.96] by ws1-6.us4.outblaze.com with http for jwest254@mail.com; Wed, 16 Jul 2003 09:30:09 -0500 From: "James West" To: "Claus Guttesen" , freebsd-performance@freebsd.org Date: Wed, 16 Jul 2003 09:30:09 -0500 X-Originating-Ip: 193.120.103.96 X-Originating-Server: ws1-6.us4.outblaze.com Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 14:30:14 -0000 No, not running IPFW or IPF or any firewall. I had MAXUSERS set to 0 but changed it to 512 when i started getting the no buffer space problems. I'm all outta ideas - could it be the intel pro nic? James ----- Original Message ----- From: Claus Guttesen Date: Wed, 16 Jul 2003 15:36:32 +0200 (CEST) To: James West , freebsd-performance@freebsd.org Subject: Re: Major problem with "No buffer space available" errors > Hi. > > > I'm having huge problems with "No buffer space > > available" errors. I've increased MAXUSERS to 512 in > > the kernel, recompiled, rebooted and the sysctl > > values below show that everything is up'ed to the > > max. > > > > Are you running ipfw on the box? When I configured a > queue without a pipe to put it on, I got the same > message and couldn't ping hosts on that interface > until I removed the queue. > > regards > Claus > > > Yahoo! Mail (http://dk.mail.yahoo.com) - Gratis: 6 MB lagerplads, spamfilter og virusscan > _______________________________________________ > freebsd-performance@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-performance > To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" > -- __________________________________________________________ Sign-up for your own FREE Personalized E-mail at Mail.com http://www.mail.com/?sr=signup CareerBuilder.com has over 400,000 jobs. Be smarter about your job search http://corp.mail.com/careers From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 07:34:32 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 73ED837B401 for ; Wed, 16 Jul 2003 07:34:32 -0700 (PDT) Received: from smtp-relay1.barrysworld.com (ns1.barrysworld.com [213.221.172.238]) by mx1.FreeBSD.org (Postfix) with ESMTP id 76D7843F75 for ; Wed, 16 Jul 2003 07:34:31 -0700 (PDT) (envelope-from killing@barrysworld.com) Received: from [213.221.181.50] (helo=barrysworld.com) by smtp-relay1.barrysworld.com with esmtp (Exim 4.20) id 19cnMG-00054y-KS for freebsd-performance@freebsd.org; Wed, 16 Jul 2003 15:34:16 +0100 Received: from vader [212.135.219.179] by barrysworld.com with ESMTP (SMTPD32-7.15) id A34263011C; Wed, 16 Jul 2003 15:37:54 +0100 Message-ID: <008a01c34ba7$533db340$b3db87d4@vader> From: "Steven Hartland" To: References: <20030716143009.20979.qmail@mail.com> Date: Wed, 16 Jul 2003 15:34:10 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: Steven Hartland List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 14:34:32 -0000 Golden rule no. 12 if you make a change and it breaks something undo the change :P Steve ----- Original Message ----- From: "James West" To: "Claus Guttesen" ; Sent: Wednesday, July 16, 2003 3:30 PM Subject: Re: Major problem with "No buffer space available" errors > No, not running IPFW or IPF or any firewall. > > I had MAXUSERS set to 0 but changed it to 512 when > i started getting the no buffer space problems. > > I'm all outta ideas - could it be the intel pro nic? > > James From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 07:44:20 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id E7DE337B401 for ; Wed, 16 Jul 2003 07:44:20 -0700 (PDT) Received: from spf13.us4.outblaze.com (205-158-62-67.outblaze.com [205.158.62.67]) by mx1.FreeBSD.org (Postfix) with ESMTP id 4C23E43F93 for ; Wed, 16 Jul 2003 07:44:20 -0700 (PDT) (envelope-from jwest254@mail.com) Received: from 205-158-62-68.outblaze.com (205-158-62-68.outblaze.com [205.158.62.68]) by spf13.us4.outblaze.com (Postfix) with QMQP id 306D3180BF33 for ; Wed, 16 Jul 2003 14:44:20 +0000 (GMT) Received: (qmail 30432 invoked from network); 16 Jul 2003 14:44:19 -0000 Received: from unknown (HELO ws1-6.us4.outblaze.com) (205.158.62.53) by 205-158-62-153.outblaze.com with SMTP; 16 Jul 2003 14:44:19 -0000 Received: (qmail 39400 invoked by uid 1001); 16 Jul 2003 14:44:19 -0000 Message-ID: <20030716144419.39399.qmail@mail.com> Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Content-Transfer-Encoding: 7bit MIME-Version: 1.0 X-Mailer: MIME-tools 5.41 (Entity 5.404) Received: from [193.120.103.96] by ws1-6.us4.outblaze.com with http for jwest254@mail.com; Wed, 16 Jul 2003 09:44:19 -0500 From: "James West" To: "Steven Hartland" , freebsd-performance@freebsd.org Date: Wed, 16 Jul 2003 09:44:19 -0500 X-Originating-Ip: 193.120.103.96 X-Originating-Server: ws1-6.us4.outblaze.com Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 14:44:21 -0000 Hmm, kinda made that all confusing sounding, what i mean is i'm having the problem with MAXUSERS set to 0 and also when its set to 512. I even cvsup'ed the source to the 4.8 code and rebuilt world & kernel and still having the same problems. James. ----- Original Message ----- From: "Steven Hartland" Date: Wed, 16 Jul 2003 15:34:10 +0100 To: Subject: Re: Major problem with "No buffer space available" errors > Golden rule no. 12 if you make a change and it breaks something undo > the change :P > > Steve > ----- Original Message ----- > From: "James West" > To: "Claus Guttesen" ; > Sent: Wednesday, July 16, 2003 3:30 PM > Subject: Re: Major problem with "No buffer space available" errors > > > > No, not running IPFW or IPF or any firewall. > > > > I had MAXUSERS set to 0 but changed it to 512 when > > i started getting the no buffer space problems. > > > > I'm all outta ideas - could it be the intel pro nic? > > > > James > > > _______________________________________________ > freebsd-performance@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-performance > To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" > -- __________________________________________________________ Sign-up for your own FREE Personalized E-mail at Mail.com http://www.mail.com/?sr=signup CareerBuilder.com has over 400,000 jobs. Be smarter about your job search http://corp.mail.com/careers From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 07:51:56 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 1F4D737B401 for ; Wed, 16 Jul 2003 07:51:56 -0700 (PDT) Received: from mailbox.univie.ac.at (mail.univie.ac.at [131.130.1.27]) by mx1.FreeBSD.org (Postfix) with ESMTP id 8706D43F75 for ; Wed, 16 Jul 2003 07:51:54 -0700 (PDT) (envelope-from l.ertl@univie.ac.at) Received: from wireless (adslle.cc.univie.ac.at [131.130.102.11]) by mailbox.univie.ac.at (8.12.2/8.12.2) with ESMTP id h6GEph3n033036; Wed, 16 Jul 2003 16:51:47 +0200 Date: Wed, 16 Jul 2003 16:51:34 +0200 (CEST) From: Lukas Ertl To: James West In-Reply-To: <20030716144419.39399.qmail@mail.com> Message-ID: <20030716164846.D609@korben.in.tern> References: <20030716144419.39399.qmail@mail.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE X-DCC-ZID-Univie-Metrics: mx1 4246; Body=0 Fuz1=0 Fuz2=0 cc: freebsd-performance@freebsd.org Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 14:51:56 -0000 On Wed, 16 Jul 2003, James West wrote: > Hmm, kinda made that all confusing sounding, what i mean is i'm having > the problem with MAXUSERS set to 0 and also when its set to 512. You didn't tell us how much memory that box has. MAXUSERS 512 is certainly a huge number. FWIW, I'd suggest you'd go back to MAXUSERS 0 (the auto-tuning of FreeBSD is quiet ok) and have a look at the mbuf clusters. You can increase them at boot time if you see you haven't enough. regards, le --=20 Lukas Ertl eMail: l.ertl@univie.ac.at UNIX-Systemadministrator Tel.: (+43 1) 4277-14073 Zentraler Informatikdienst (ZID) Fax.: (+43 1) 4277-9140 der Universit=E4t Wien http://mailbox.univie.ac.at/~le/ From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 08:00:19 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 0EEC637B401 for ; Wed, 16 Jul 2003 08:00:19 -0700 (PDT) Received: from otter3.centtech.com (moat3.centtech.com [207.200.51.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 59C2D43F85 for ; Wed, 16 Jul 2003 08:00:18 -0700 (PDT) (envelope-from anderson@centtech.com) Received: from centtech.com (dhcp-243.centtech.com [204.177.173.243]) by otter3.centtech.com (8.12.3/8.12.3) with ESMTP id h6GF0H56090450; Wed, 16 Jul 2003 10:00:17 -0500 (CDT) (envelope-from anderson@centtech.com) Message-ID: <3F15686F.3020300@centtech.com> Date: Wed, 16 Jul 2003 09:59:59 -0500 From: Eric Anderson User-Agent: Mozilla/5.0 (X11; U; Linux i386; en-US; rv:1.0.1) Gecko/20020823 Netscape/7.0 X-Accept-Language: en-us, en MIME-Version: 1.0 To: James West References: <20030716125123.53307.qmail@mail.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 15:00:19 -0000 James West wrote: > Hi, > > I'm having huge problems with "No buffer space available" errors. I've increased MAXUSERS to 512 in the kernel, recompiled, rebooted and the sysctl values below show that everything is up'ed to the max. > > some details which may help are: > > kern.maxfiles: 10796 > kern.maxfilesperproc: 9716 > kern.ipc.maxsockbuf: 262144 > kern.ipc.somaxconn: 65000 > kern.ipc.nmbclusters: 65000 > kern.ipc.maxsockets: 65000 Try reducing the somaxconn and maxsockets. How much memory do you have on these machines? Eric -- ------------------------------------------------------------------ Eric Anderson Systems Administrator Centaur Technology Attitudes are contagious, is yours worth catching? ------------------------------------------------------------------ From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 08:19:15 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 74C7E37B401 for ; Wed, 16 Jul 2003 08:19:15 -0700 (PDT) Received: from spf13.us4.outblaze.com (205-158-62-67.outblaze.com [205.158.62.67]) by mx1.FreeBSD.org (Postfix) with ESMTP id 278F943FBF for ; Wed, 16 Jul 2003 08:19:12 -0700 (PDT) (envelope-from jwest254@mail.com) Received: from 205-158-62-68.outblaze.com (205-158-62-68.outblaze.com [205.158.62.68]) by spf13.us4.outblaze.com (Postfix) with QMQP id 966B31814A33 for ; Wed, 16 Jul 2003 15:19:11 +0000 (GMT) Received: (qmail 43675 invoked from network); 16 Jul 2003 15:19:06 -0000 Received: from unknown (HELO ws1-7.us4.outblaze.com) (205.158.62.57) by 205-158-62-153.outblaze.com with SMTP; 16 Jul 2003 15:19:06 -0000 Received: (qmail 39092 invoked by uid 1001); 16 Jul 2003 15:19:05 -0000 Message-ID: <20030716151905.39091.qmail@mail.com> Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Content-Transfer-Encoding: 7bit MIME-Version: 1.0 X-Mailer: MIME-tools 5.41 (Entity 5.404) Received: from [193.120.103.96] by ws1-7.us4.outblaze.com with http for jwest254@mail.com; Wed, 16 Jul 2003 10:19:05 -0500 From: "James West" To: "Lukas Ertl" Date: Wed, 16 Jul 2003 10:19:05 -0500 X-Originating-Ip: 193.120.103.96 X-Originating-Server: ws1-7.us4.outblaze.com cc: freebsd-performance@freebsd.org Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 15:19:15 -0000 The machine has 256Mb of RAM, 36Gb scsi drive and is a 1.3Ghz AMD cpu. I've switched MAXUSERS back to 0 now, right now before reboot my mbuf sysctl values read: # sysctl -a | grep mbuf kern.ipc.mbuf_wait: 32 kern.ipc.nmbufs: 260000 I'm stumped, stumped like a tree! James. ----- Original Message ----- From: Lukas Ertl Date: Wed, 16 Jul 2003 16:51:34 +0200 (CEST) To: James West Subject: Re: Major problem with "No buffer space available" errors > On Wed, 16 Jul 2003, James West wrote: > > > Hmm, kinda made that all confusing sounding, what i mean is i'm having > > the problem with MAXUSERS set to 0 and also when its set to 512. > > You didn't tell us how much memory that box has. MAXUSERS 512 is certainly > a huge number. > > FWIW, I'd suggest you'd go back to MAXUSERS 0 (the auto-tuning of FreeBSD > is quiet ok) and have a look at the mbuf clusters. You can increase them > at boot time if you see you haven't enough. > > regards, > le > > -- > Lukas Ertl eMail: l.ertl@univie.ac.at > UNIX-Systemadministrator Tel.: (+43 1) 4277-14073 > Zentraler Informatikdienst (ZID) Fax.: (+43 1) 4277-9140 > der Universität Wien http://mailbox.univie.ac.at/~le/ > _______________________________________________ > freebsd-performance@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-performance > To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" > -- __________________________________________________________ Sign-up for your own FREE Personalized E-mail at Mail.com http://www.mail.com/?sr=signup CareerBuilder.com has over 400,000 jobs. Be smarter about your job search http://corp.mail.com/careers From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 09:39:19 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id EC2F637B401 for ; Wed, 16 Jul 2003 09:39:19 -0700 (PDT) Received: from adsl-63-198-35-122.dsl.snfc21.pacbell.net (adsl-63-198-35-122.dsl.snfc21.pacbell.net [63.198.35.122]) by mx1.FreeBSD.org (Postfix) with ESMTP id 01A2E43F3F for ; Wed, 16 Jul 2003 09:39:19 -0700 (PDT) (envelope-from j_guojun@lbl.gov) Received: from lbl.gov (localhost.pacbell.net [127.0.0.1]) ESMTP id h6GGfCFA000364; Wed, 16 Jul 2003 09:41:12 -0700 (PDT) (envelope-from j_guojun@lbl.gov) Sender: jin@adsl-63-198-35-122.dsl.snfc21.pacbell.net Message-ID: <3F158028.9A91EE54@lbl.gov> Date: Wed, 16 Jul 2003 09:41:12 -0700 From: "Jin Guojun [NCS]" X-Mailer: Mozilla 4.76 [en] (X11; U; FreeBSD 4.8-RELEASE i386) X-Accept-Language: zh, zh-CN, en-US, en MIME-Version: 1.0 To: James West References: <20030716151905.39091.qmail@mail.com> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit cc: freebsd-performance@freebsd.org cc: Lukas Ertl Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 16:39:20 -0000 Ok, this may have confirmed that it is a drive or motherboard (Intel specific) bug. We had same problem about a month ago on Intel D875PBZP motherboard with 5.1-RELEASE. It has a onboard Intel 10/100/1000 PRO (em) NIC. We left the machine idle for a month, since the machine is useless and we have plentiful other things to do. This week, we decide to use this machine for high-speed network test. Somehow, the machine just works without any change. I asked around and no one here says that they changed any thing. What I can suggest to you is to test a different NIC to see if the problem persists. If it is, then there is a kernel problem, otherwise, it is Intel problem. -Jin James West wrote: > The machine has 256Mb of RAM, 36Gb scsi drive and is a 1.3Ghz AMD cpu. > > I've switched MAXUSERS back to 0 now, right now before reboot my mbuf sysctl values read: > > # sysctl -a | grep mbuf > kern.ipc.mbuf_wait: 32 > kern.ipc.nmbufs: 260000 > > I'm stumped, stumped like a tree! > > James. > > ----- Original Message ----- > From: Lukas Ertl > Date: Wed, 16 Jul 2003 16:51:34 +0200 (CEST) > To: James West > Subject: Re: Major problem with "No buffer space available" errors > > > On Wed, 16 Jul 2003, James West wrote: > > > > > Hmm, kinda made that all confusing sounding, what i mean is i'm having > > > the problem with MAXUSERS set to 0 and also when its set to 512. > > > > You didn't tell us how much memory that box has. MAXUSERS 512 is certainly > > a huge number. > > > > FWIW, I'd suggest you'd go back to MAXUSERS 0 (the auto-tuning of FreeBSD > > is quiet ok) and have a look at the mbuf clusters. You can increase them > > at boot time if you see you haven't enough. > > > > regards, > > le > > > > -- > > Lukas Ertl eMail: l.ertl@univie.ac.at > > UNIX-Systemadministrator Tel.: (+43 1) 4277-14073 > > Zentraler Informatikdienst (ZID) Fax.: (+43 1) 4277-9140 > > der Universität Wien http://mailbox.univie.ac.at/~le/ > > _______________________________________________ > > freebsd-performance@freebsd.org mailing list > > http://lists.freebsd.org/mailman/listinfo/freebsd-performance > > To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" > > > > -- > __________________________________________________________ > Sign-up for your own FREE Personalized E-mail at Mail.com > http://www.mail.com/?sr=signup > > CareerBuilder.com has over 400,000 jobs. Be smarter about your job search > http://corp.mail.com/careers > > _______________________________________________ > freebsd-performance@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-performance > To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 09:59:50 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 360DF37B401 for ; Wed, 16 Jul 2003 09:59:50 -0700 (PDT) Received: from godel.mtl.distributel.net (nat.MTL.distributel.NET [66.38.181.24]) by mx1.FreeBSD.org (Postfix) with ESMTP id 6B2F743F93 for ; Wed, 16 Jul 2003 09:59:49 -0700 (PDT) (envelope-from bmilekic@technokratis.com) Received: from godel.mtl.distributel.net (localhost [127.0.0.1]) h6GD3sEH014872; Wed, 16 Jul 2003 13:03:54 GMT (envelope-from bmilekic@technokratis.com) Received: (from bmilekic@localhost) by godel.mtl.distributel.net (8.12.9/8.12.9/Submit) id h6GD3st2014871; Wed, 16 Jul 2003 13:03:54 GMT X-Authentication-Warning: godel.mtl.distributel.net: bmilekic set sender to bmilekic@technokratis.com using -f Date: Wed, 16 Jul 2003 13:03:54 +0000 From: Bosko Milekic To: James West Message-ID: <20030716130354.GA14831@technokratis.com> References: <20030716151905.39091.qmail@mail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20030716151905.39091.qmail@mail.com> User-Agent: Mutt/1.4.1i cc: freebsd-performance@freebsd.org Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 16:59:50 -0000 On Wed, Jul 16, 2003 at 10:19:05AM -0500, James West wrote: [... ... ... ...] Just show us 'netstat -m' so that we can confirm that this probably isn't a nmbclusters issue. When it starts happening, what happens if you ifconfig down the net if and then bring it back up? -- Bosko Milekic * bmilekic@technokratis.com * bmilekic@FreeBSD.org TECHNOkRATIS Consulting Services * http://www.technokratis.com/ From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 11:06:04 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id CF0A037B401 for ; Wed, 16 Jul 2003 11:06:04 -0700 (PDT) Received: from wingerboy.noc.sonic.net (wingerboy.noc.sonic.net [64.142.18.2]) by mx1.FreeBSD.org (Postfix) with ESMTP id 5554143FA3 for ; Wed, 16 Jul 2003 11:06:04 -0700 (PDT) (envelope-from kgc@wingerboy.noc.sonic.net) Received: from wingerboy.noc.sonic.net (localhost.sonic.net [127.0.0.1]) h6GI64Og044710 for ; Wed, 16 Jul 2003 11:06:04 -0700 (PDT) (envelope-from kgc@wingerboy.noc.sonic.net) Received: (from kgc@localhost) by wingerboy.noc.sonic.net (8.12.9/8.12.7/Submit) id h6GI63h7044709 for freebsd-performance@freebsd.org; Wed, 16 Jul 2003 11:06:04 -0700 (PDT) Date: Wed, 16 Jul 2003 11:06:03 -0700 From: Kelsey Cummings To: freebsd-performance@freebsd.org Message-ID: <20030716180603.GM95278@sonic.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.4.1i X-Mailman-Approved-At: Wed, 16 Jul 2003 11:07:37 -0700 Subject: Qlogic 2200 performance issues? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 18:06:05 -0000 We've got a 5.1 box that we are using to play around with vinum and fcal. It's a dual PIII-800, gig of ram, etc. The Qlogic 2200 has 2 old NetApp FC8 shelves on them, with 14 disks. Each one of these disks is able to do a sustained 17MBs in sequential IO. However, when all disks are accessed at once sustained I/O drops to ~1-4MBs per device, or ~28MBs total for all devices together. We see the same problem with vinum, stipping all 14 disks together results in ~28MBs, which is the same as what's gotten with 2, 4, 6, 8, .. disks. During these tests the box is responsive, CPU ultilization is minimal, etc. We don't have much experience with fcal here but this doesn't make alot of sense to us. It would appear that the controller or loop is getting staturated at only 28MBs? Can anyone shed some light? We've just be using dd (single and multiple processes) to test the sequential IO throughput. -- Kelsey Cummings - kgc@sonic.net sonic.net System Administrator 2260 Apollo Way 707.522.1000 (Voice) Santa Rosa, CA 95407 707.547.2199 (Fax) http://www.sonic.net/ Fingerprint = 7F 59 43 1B 44 8A 0D 57 91 08 73 73 7A 48 90 C5 From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 14:37:16 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id D7A0A37B401 for ; Wed, 16 Jul 2003 14:37:16 -0700 (PDT) Received: from mta1.lbl.gov (mta1.lbl.gov [128.3.41.24]) by mx1.FreeBSD.org (Postfix) with ESMTP id 272DE43FA3 for ; Wed, 16 Jul 2003 14:37:16 -0700 (PDT) (envelope-from j_guojun@lbl.gov) Received: from mta1.lbl.gov (localhost [127.0.0.1]) by mta1.lbl.gov (8.12.9/8.12.9) with ESMTP id h6GLbDlq000113 for ; Wed, 16 Jul 2003 14:37:14 -0700 (PDT) Received: from lbl.gov (gracie.lbl.gov [131.243.2.175]) by mta1.lbl.gov (8.12.9/8.12.9) with ESMTP id h6GLbD3X000110 for ; Wed, 16 Jul 2003 14:37:13 -0700 (PDT) Sender: jin@lbl.gov Message-ID: <3F15C584.38848DCE@lbl.gov> Date: Wed, 16 Jul 2003 14:37:08 -0700 From: "Jin Guojun [DSD]" X-Mailer: Mozilla 4.76 [en] (X11; U; FreeBSD 4.8-RELEASE i386) X-Accept-Language: zh, zh-CN, en MIME-Version: 1.0 To: freebsd-performance@freebsd.org Content-Type: multipart/mixed; boundary="------------AD7BA6D51EB6C406023AC363" X-Content-Filtered-By: Mailman/MimeDel 2.1.1 Subject: [Fwd: timer counter chip access mystery] X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 21:37:17 -0000 This is a multi-part message in MIME format. --------------AD7BA6D51EB6C406023AC363 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit i386/isa/clock.c, line 1207 i8254_get_timecount(...) { outb(TIMER_MODE, TIMER_SEL0 | TIMER_LATCH); low = inb(TIMER_CNTR0); high = inb(TIMER_CNTR0); ... inb(IO_ICU1); return counter. } This routine takes about 4000 ns. It makes gettimeofday() cost over 4000 ns. measure_i8254_get_timecount: Two_TIMER_CNTR0 2362 ns measure_i8254_get_timecount: IO_ICU1 936 ns and outb() looks like to take another 700 ns. The Linux uses the same process to get the time counter (see below). It also comments that this is from Steve McCanne's microtime-i386 for BSD. However, gettimeofday() under Linux costs 900 ns. See attached GIF file for a table of comparison. Any idea why linux is 4 times faster than FreeBSD in reading the same time counter chip? -Jin ----------------- Linux source code for the same routine ----------- arch/x86_64/kernel/time.c (line 68) void do_gettimeofday(struct timeval *tv) { unsigned long flags, t; unsigned int sec, usec; read_lock_irqsave(&xtime_lock, flags); spin_lock(&time_offset_lock); sec = xtime.tv_sec; usec = xtime.tv_usec; t = (jiffies - wall_jiffies) * (1000000L / HZ) + do_gettimeoffset(); if (t > timeoffset) timeoffset = t; usec += timeoffset; spin_unlock(&time_offset_lock); read_unlock_irqrestore(&xtime_lock, flags); tv->tv_sec = sec + usec / 1000000; tv->tv_usec = usec % 1000000; } arch/i386/kernel/time.c (line 127) /* This function must be called with interrupts disabled * It was inspired by Steve McCanne's microtime-i386 for BSD. -- jrs * * However, the pc-audio speaker driver changes the divisor so that * it gets interrupted rather more often - it loads 64 into the * counter rather than 11932! This has an adverse impact on * do_gettimeoffset() -- it stops working! What is also not * good is that the interval that our timer function gets called * is no longer 10.0002 ms, but 9.9767 ms. To get around this * would require using a different timing source. Maybe someone * could use the RTC - I know that this can interrupt at frequencies * ranging from 8192Hz to 2Hz. If I had the energy, I'd somehow fix * it so that at startup, the timer code in sched.c would select * using either the RTC or the 8253 timer. The decision would be * based on whether there was any other device around that needed * to trample on the 8253. I'd set up the RTC to interrupt at 1024 Hz, * and then do some jiggery to have a version of do_timer that * advanced the clock by 1/1024 s. Every time that reached over 1/100 * of a second, then do all the old code. If the time was kept correct * then do_gettimeoffset could just return 0 - there is no low order * divider that can be accessed. * * Ideally, you would be able to use the RTC for the speaker driver, * but it appears that the speaker driver really needs interrupt more * often than every 120 us or so. * * Anyway, this needs more thought.... pjsg (1993-08-28) * * If you are really that interested, you should be reading * comp.protocols.time.ntp! */ static unsigned long do_slow_gettimeoffset(void) { int count; static int count_p = LATCH; /* for the first call after boot */ static unsigned long jiffies_p = 0; /* * cache volatile jiffies temporarily; we have IRQs turned off. */ unsigned long jiffies_t; /* gets recalled with irq locally disabled */ spin_lock(&i8253_lock); /* timer count may underflow right here */ outb_p(0x00, 0x43); /* latch the count ASAP */ count = inb_p(0x40); /* read the latched count */ jiffies_t = jiffies; count |= inb_p(0x40) << 8; /* VIA686a test code... reset the latch if count > max + 1 */ if (count > LATCH) { outb_p(0x34, 0x43); outb_p(LATCH & 0xff, 0x40); outb(LATCH >> 8, 0x40); count = LATCH - 1; } spin_unlock(&i8253_lock); ... return count; } --------------AD7BA6D51EB6C406023AC363-- From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 14:44:41 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 4582C37B401 for ; Wed, 16 Jul 2003 14:44:41 -0700 (PDT) Received: from silver.he.iki.fi (silver.he.iki.fi [193.64.42.241]) by mx1.FreeBSD.org (Postfix) with ESMTP id 249EB43F75 for ; Wed, 16 Jul 2003 14:44:40 -0700 (PDT) (envelope-from pete@he.iki.fi) Received: from PETEX31 (h81.vuokselantie10.fi [193.64.42.129]) by silver.he.iki.fi (8.12.9/8.11.4) with SMTP id h6GLicsL010061; Thu, 17 Jul 2003 00:44:38 +0300 (EEST) (envelope-from pete@he.iki.fi) Message-ID: <071401c34be3$71a03510$812a40c1@PETEX31> From: "Petri Helenius" To: "Jin Guojun [DSD]" , References: <3F15C584.38848DCE@lbl.gov> Date: Thu, 17 Jul 2003 00:44:32 +0300 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Subject: Re: timer counter chip access mystery] X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 21:44:41 -0000 This access happens over ISA bus and thus happens at the speed the bus operates at. Use TSC or ACPI for faster gettimeofday. Pete ----- Original Message ----- From: "Jin Guojun [DSD]" To: Sent: Thursday, July 17, 2003 12:37 AM Subject: [Fwd: timer counter chip access mystery] > i386/isa/clock.c, line 1207 > i8254_get_timecount(...) > { > outb(TIMER_MODE, TIMER_SEL0 | TIMER_LATCH); > > low = inb(TIMER_CNTR0); > high = inb(TIMER_CNTR0); > ... > inb(IO_ICU1); > return counter. > } > > This routine takes about 4000 ns. It makes gettimeofday() > cost over 4000 ns. > > measure_i8254_get_timecount: Two_TIMER_CNTR0 2362 ns > measure_i8254_get_timecount: IO_ICU1 936 ns > and outb() looks like to take another 700 ns. > > The Linux uses the same process to get the time counter (see below). > It also comments that this is from Steve McCanne's microtime-i386 for BSD. > However, gettimeofday() under Linux costs 900 ns. > See attached GIF file for a table of comparison. > > Any idea why linux is 4 times faster than FreeBSD in > reading the same time counter chip? > > -Jin > > ----------------- Linux source code for the same routine ----------- > > arch/x86_64/kernel/time.c (line 68) > void do_gettimeofday(struct timeval *tv) > { > unsigned long flags, t; > unsigned int sec, usec; > > read_lock_irqsave(&xtime_lock, flags); > spin_lock(&time_offset_lock); > > sec = xtime.tv_sec; > usec = xtime.tv_usec; > > t = (jiffies - wall_jiffies) * (1000000L / HZ) + do_gettimeoffset(); > if (t > timeoffset) timeoffset = t; > usec += timeoffset; > > spin_unlock(&time_offset_lock); > read_unlock_irqrestore(&xtime_lock, flags); > > tv->tv_sec = sec + usec / 1000000; > tv->tv_usec = usec % 1000000; > } > > arch/i386/kernel/time.c (line 127) > > /* This function must be called with interrupts disabled > * It was inspired by Steve McCanne's microtime-i386 for BSD. -- jrs > * > * However, the pc-audio speaker driver changes the divisor so that > * it gets interrupted rather more often - it loads 64 into the > * counter rather than 11932! This has an adverse impact on > * do_gettimeoffset() -- it stops working! What is also not > * good is that the interval that our timer function gets called > * is no longer 10.0002 ms, but 9.9767 ms. To get around this > * would require using a different timing source. Maybe someone > * could use the RTC - I know that this can interrupt at frequencies > * ranging from 8192Hz to 2Hz. If I had the energy, I'd somehow fix > * it so that at startup, the timer code in sched.c would select > * using either the RTC or the 8253 timer. The decision would be > * based on whether there was any other device around that needed > * to trample on the 8253. I'd set up the RTC to interrupt at 1024 Hz, > * and then do some jiggery to have a version of do_timer that > * advanced the clock by 1/1024 s. Every time that reached over 1/100 > * of a second, then do all the old code. If the time was kept correct > * then do_gettimeoffset could just return 0 - there is no low order > * divider that can be accessed. > * > * Ideally, you would be able to use the RTC for the speaker driver, > * but it appears that the speaker driver really needs interrupt more > * often than every 120 us or so. > * > * Anyway, this needs more thought.... pjsg (1993-08-28) > * > * If you are really that interested, you should be reading > * comp.protocols.time.ntp! > */ > > static unsigned long do_slow_gettimeoffset(void) > { > int count; > > static int count_p = LATCH; /* for the first call after boot */ > static unsigned long jiffies_p = 0; > > /* > * cache volatile jiffies temporarily; we have IRQs turned off. > */ > unsigned long jiffies_t; > > /* gets recalled with irq locally disabled */ > spin_lock(&i8253_lock); > /* timer count may underflow right here */ > outb_p(0x00, 0x43); /* latch the count ASAP */ > > count = inb_p(0x40); /* read the latched count */ > > jiffies_t = jiffies; > > count |= inb_p(0x40) << 8; > > /* VIA686a test code... reset the latch if count > max + 1 */ > if (count > LATCH) { > outb_p(0x34, 0x43); > outb_p(LATCH & 0xff, 0x40); > outb(LATCH >> 8, 0x40); > count = LATCH - 1; > } > > spin_unlock(&i8253_lock); > ... > return count; > } -------------------------------------------------------------------------------- > _______________________________________________ > freebsd-performance@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-performance > To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" > From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 14:58:50 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 4EF9037B40D for ; Wed, 16 Jul 2003 14:58:49 -0700 (PDT) Received: from smtp-relay2.barrysworld.com (smtp-relay2.barrysworld.com [213.221.172.236]) by mx1.FreeBSD.org (Postfix) with ESMTP id 7B02243F75 for ; Wed, 16 Jul 2003 14:58:48 -0700 (PDT) (envelope-from killing@barrysworld.com) Received: from [213.221.181.50] (helo=barrysworld.com) by smtp-relay2.barrysworld.com with esmtp (Exim 4.20) id 19cuGw-0007tU-Ab; Wed, 16 Jul 2003 22:57:14 +0100 Received: from vader [212.135.219.179] by barrysworld.com with ESMTP (SMTPD32-7.15) id AB14531100D6; Wed, 16 Jul 2003 23:00:52 +0100 Message-ID: <00b501c34be5$351a83f0$b3db87d4@vader> From: "Steven Hartland" To: "Petri Helenius" , "Jin Guojun [DSD]" , References: <3F15C584.38848DCE@lbl.gov> <071401c34be3$71a03510$812a40c1@PETEX31> Date: Wed, 16 Jul 2003 22:57:07 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Subject: Re: timer counter chip access mystery] X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: Steven Hartland List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 21:58:51 -0000 If this could be changed / fixed it would have a marked performance increase for all game server code which use this function heavily from what I've seen and would explain the sometimes huge performance differential between FreeBSD and Linux. Steve / K ----- Original Message ----- From: "Petri Helenius" To: "Jin Guojun [DSD]" ; Sent: Wednesday, July 16, 2003 10:44 PM Subject: Re: timer counter chip access mystery] > > This access happens over ISA bus and thus happens at the speed the bus > operates at. Use TSC or ACPI for faster gettimeofday. > > Pete From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 15:01:57 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id A9F9537B401 for ; Wed, 16 Jul 2003 15:01:57 -0700 (PDT) Received: from silver.he.iki.fi (silver.he.iki.fi [193.64.42.241]) by mx1.FreeBSD.org (Postfix) with ESMTP id A26CD43F3F for ; Wed, 16 Jul 2003 15:01:56 -0700 (PDT) (envelope-from pete@he.iki.fi) Received: from PETEX31 (h81.vuokselantie10.fi [193.64.42.129]) by silver.he.iki.fi (8.12.9/8.11.4) with SMTP id h6GM1tsL010222; Thu, 17 Jul 2003 01:01:55 +0300 (EEST) (envelope-from pete@he.iki.fi) Message-ID: <075101c34be5$db5e4c10$812a40c1@PETEX31> From: "Petri Helenius" To: "Steven Hartland" , "Jin Guojun [DSD]" , References: <3F15C584.38848DCE@lbl.gov> <071401c34be3$71a03510$812a40c1@PETEX31> <00b501c34be5$351a83f0$b3db87d4@vader> Date: Thu, 17 Jul 2003 01:01:49 +0300 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Subject: Re: timer counter chip access mystery] X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 22:01:58 -0000 5.x defaults to ACPI or TSC timers. So your wish has been transported by a timemachine into the past. With 4.x you have to set it manually. Pete ----- Original Message ----- From: "Steven Hartland" To: "Petri Helenius" ; "Jin Guojun [DSD]" ; Sent: Thursday, July 17, 2003 12:57 AM Subject: Re: timer counter chip access mystery] > If this could be changed / fixed it would have a marked performance increase > for all game server code which use this function heavily from what I've seen > and would explain the sometimes huge performance differential between > FreeBSD and Linux. > > Steve / K > ----- Original Message ----- > From: "Petri Helenius" > To: "Jin Guojun [DSD]" ; > Sent: Wednesday, July 16, 2003 10:44 PM > Subject: Re: timer counter chip access mystery] > > > > > > This access happens over ISA bus and thus happens at the speed the bus > > operates at. Use TSC or ACPI for faster gettimeofday. > > > > Pete > > From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 15:07:51 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 51EBE37B404 for ; Wed, 16 Jul 2003 15:07:51 -0700 (PDT) Received: from ws1-13.us4.outblaze.com (205-158-62-112.outblaze.com [205.158.62.112]) by mx1.FreeBSD.org (Postfix) with SMTP id 424DF43F93 for ; Wed, 16 Jul 2003 15:07:48 -0700 (PDT) (envelope-from jwest254@mail.com) Received: (qmail 21955 invoked by uid 1001); 16 Jul 2003 22:06:22 -0000 Message-ID: <20030716220622.21950.qmail@mail.com> Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Content-Transfer-Encoding: 7bit MIME-Version: 1.0 X-Mailer: MIME-tools 5.41 (Entity 5.404) Received: from [194.165.173.182] by ws1-88.us4.outblaze.com with http for jwest254@mail.com; Wed, 16 Jul 2003 17:06:22 -0500 From: "James West" To: "Bosko Milekic" Date: Wed, 16 Jul 2003 17:06:22 -0500 X-Originating-Ip: 194.165.173.182 X-Originating-Server: ws1-88.us4.outblaze.com cc: freebsd-performance@freebsd.org Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 22:07:51 -0000 73/432/260000 mbufs in use (current/peak/max): 68 mbufs allocated to data 5 mbufs allocated to packet headers 64/230/65000 mbuf clusters in use (current/peak/max) 568 Kbytes allocated to network (0% of mb_map in use) 0 requests for memory denied 0 requests for memory delayed 0 calls to protocol drain routines ----- Original Message ----- From: Bosko Milekic Date: Wed, 16 Jul 2003 13:03:54 +0000 To: James West Subject: Re: Major problem with "No buffer space available" errors > > On Wed, Jul 16, 2003 at 10:19:05AM -0500, James West wrote: > [... ... ... ...] > > Just show us 'netstat -m' so that we can confirm that this probably > isn't a nmbclusters issue. > > When it starts happening, what happens if you ifconfig down the net if > and then bring it back up? > > -- > Bosko Milekic * bmilekic@technokratis.com * bmilekic@FreeBSD.org > TECHNOkRATIS Consulting Services * http://www.technokratis.com/ -- __________________________________________________________ Sign-up for your own FREE Personalized E-mail at Mail.com http://www.mail.com/?sr=signup CareerBuilder.com has over 400,000 jobs. Be smarter about your job search http://corp.mail.com/careers From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 15:26:45 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 764AA37B401 for ; Wed, 16 Jul 2003 15:26:45 -0700 (PDT) Received: from smtp-relay2.barrysworld.com (smtp-relay2.barrysworld.com [213.221.172.236]) by mx1.FreeBSD.org (Postfix) with ESMTP id C9BB143F3F for ; Wed, 16 Jul 2003 15:26:44 -0700 (PDT) (envelope-from killing@barrysworld.com) Received: from [213.221.181.50] (helo=barrysworld.com) by smtp-relay2.barrysworld.com with esmtp (Exim 4.20) id 19cujM-0008F1-RF; Wed, 16 Jul 2003 23:26:36 +0100 Received: from vader [212.135.219.179] by barrysworld.com with ESMTP (SMTPD32-7.15) id A1F6580126; Wed, 16 Jul 2003 23:30:14 +0100 Message-ID: <00da01c34be9$4f6702c0$b3db87d4@vader> From: "Steven Hartland" To: "Petri Helenius" , "Jin Guojun [DSD]" , References: <3F15C584.38848DCE@lbl.gov> <071401c34be3$71a03510$812a40c1@PETEX31> <00b501c34be5$351a83f0$b3db87d4@vader> <075101c34be5$db5e4c10$812a40c1@PETEX31> Date: Wed, 16 Jul 2003 23:26:32 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Subject: Re: timer counter chip access mystery] X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: Steven Hartland List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 22:26:45 -0000 Ok have to look else where for the differential as this was 5.0 and 5.1 :( Steve ----- Original Message ----- From: "Petri Helenius" To: "Steven Hartland" ; "Jin Guojun [DSD]" ; Sent: Wednesday, July 16, 2003 11:01 PM Subject: Re: timer counter chip access mystery] > > 5.x defaults to ACPI or TSC timers. So your wish has been transported > by a timemachine into the past. > > With 4.x you have to set it manually. > > Pete From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 16:23:17 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id DDF9237B401 for ; Wed, 16 Jul 2003 16:23:17 -0700 (PDT) Received: from mta1.lbl.gov (mta1.lbl.gov [128.3.41.24]) by mx1.FreeBSD.org (Postfix) with ESMTP id 277CD43F85 for ; Wed, 16 Jul 2003 16:23:17 -0700 (PDT) (envelope-from j_guojun@lbl.gov) Received: from mta1.lbl.gov (localhost [127.0.0.1]) by mta1.lbl.gov (8.12.9/8.12.9) with ESMTP id h6GNN9lu022621 for ; Wed, 16 Jul 2003 16:23:14 -0700 (PDT) Received: from lbl.gov (gracie.lbl.gov [131.243.2.175]) by mta1.lbl.gov (8.12.9/8.12.9) with ESMTP id h6GNN43X022609; Wed, 16 Jul 2003 16:23:09 -0700 (PDT) Sender: jin@lbl.gov Message-ID: <3F15DE52.130D8FA9@lbl.gov> Date: Wed, 16 Jul 2003 16:22:58 -0700 From: "Jin Guojun [DSD]" X-Mailer: Mozilla 4.76 [en] (X11; U; FreeBSD 4.8-RELEASE i386) X-Accept-Language: zh, zh-CN, en MIME-Version: 1.0 To: Petri Helenius References: <3F15C584.38848DCE@lbl.gov> <071401c34be3$71a03510$812a40c1@PETEX31> <00b501c34be5$351a83f0$b3db87d4@vader> <075101c34be5$db5e4c10$812a40c1@PETEX31> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org cc: Steven Hartland Subject: Re: timer counter chip access mystery] X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Jul 2003 23:23:18 -0000 5.1 defaults to use ACPI-safe, which calls read_counter() at least three times. The total cost is 3098 ns. So, it meas that read_count() cost is about 1000 ns, which is close to what the Linux does. The rest question is how much error will be if not to use ACPI-safe? i.e., sysctl kern.timecounter.hardware=ACPI? -Jin /* * Fetch current time value from reliable hardware. */ static unsigned acpi_timer_get_timecount(struct timecounter *tc) { return (read_counter()); } /* * Fetch current time value from hardware that may not correctly * latch the counter. */ static unsigned acpi_timer_get_timecount_safe(struct timecounter *tc) { unsigned u1, u2, u3; u2 = read_counter(); u3 = read_counter(); do { u1 = u2; u2 = u3; u3 = read_counter(); } while (u1 > u2 || u2 > u3 || (u3 - u1) > 15); return (u2); } Petri Helenius wrote: > 5.x defaults to ACPI or TSC timers. So your wish has been transported > by a timemachine into the past. > > With 4.x you have to set it manually. > > Pete > > ----- Original Message ----- > From: "Steven Hartland" > To: "Petri Helenius" ; "Jin Guojun [DSD]" ; > Sent: Thursday, July 17, 2003 12:57 AM > Subject: Re: timer counter chip access mystery] > > > If this could be changed / fixed it would have a marked performance increase > > for all game server code which use this function heavily from what I've seen > > and would explain the sometimes huge performance differential between > > FreeBSD and Linux. > > > > Steve / K > > ----- Original Message ----- > > From: "Petri Helenius" > > To: "Jin Guojun [DSD]" ; > > Sent: Wednesday, July 16, 2003 10:44 PM > > Subject: Re: timer counter chip access mystery] > > > > > > > > > > This access happens over ISA bus and thus happens at the speed the bus > > > operates at. Use TSC or ACPI for faster gettimeofday. > > > > > > Pete From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 22:42:45 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id E54C137B401 for ; Wed, 16 Jul 2003 22:42:45 -0700 (PDT) Received: from puffin.mail.pas.earthlink.net (puffin.mail.pas.earthlink.net [207.217.120.139]) by mx1.FreeBSD.org (Postfix) with ESMTP id 5617443F93 for ; Wed, 16 Jul 2003 22:42:45 -0700 (PDT) (envelope-from tlambert2@mindspring.com) Received: from user-uinj8ql.dialup.mindspring.com ([165.121.163.85] helo=mindspring.com) by puffin.mail.pas.earthlink.net with asmtp (SSLv3:RC4-MD5:128) (Exim 3.33 #1) id 19d1XD-0000lf-00; Wed, 16 Jul 2003 22:42:32 -0700 Message-ID: <3F16370C.674EE08D@mindspring.com> Date: Wed, 16 Jul 2003 22:41:32 -0700 From: Terry Lambert X-Mailer: Mozilla 4.79 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: James West References: <20030716125123.53307.qmail@mail.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-ELNK-Trace: b1a02af9316fbb217a47c185c03b154d40683398e744b8a4d1da5842b0fa92832ffe02efbbca9093350badd9bab72f9c350badd9bab72f9c350badd9bab72f9c cc: freebsd-performance@freebsd.org Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 17 Jul 2003 05:42:46 -0000 James West wrote: > I'm having huge problems with "No buffer space available" errors. I've > increased MAXUSERS to 512 in the kernel, recompiled, rebooted and the > sysctl values below show that everything is up'ed to the max. Historically, the most common cause of this message has been an interface you were actively feeding data going down. Most commonly, this occurs in the use of UDP and ICMP protocols (e.g. "ping") out a downed interface, since they are not constrained by your sendspace settings, which apply to TCP (they will basically permit you to use all your buffer in the attempt). So the questions to answer are: 1) Are you using a lot of ICMP (e.g. ping, traceroute, RIP, etc.)? 2) Are you using a lot of UDP (e.g. Linux NFS clients using UDP mounts and an rsize or wsize larger than the MTU would permit to fit in a single UDP packet)? Other than that, you should "netstat -an" and add up the contents of the SendQ/RecvQ columns. It's possible that, in fact, you *are* running out of buffer space. For the default of 64K, you would need 2G of RAM dedicated to nothing but mbufs (not including headers!) to support only 32,768 simultaneous connections without mbuf overcommit. Dropping the sndspace/rcvspace paramaters back to their pre-bump-up defaults will double the number of connections for the same amount of RAM. -- Terry From owner-freebsd-performance@FreeBSD.ORG Wed Jul 16 23:51:34 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 67B5B37B401 for ; Wed, 16 Jul 2003 23:51:34 -0700 (PDT) Received: from silver.he.iki.fi (silver.he.iki.fi [193.64.42.241]) by mx1.FreeBSD.org (Postfix) with ESMTP id 49D9143F3F for ; Wed, 16 Jul 2003 23:51:30 -0700 (PDT) (envelope-from pete@he.iki.fi) Received: from PETEX31 (h81.vuokselantie10.fi [193.64.42.129]) by silver.he.iki.fi (8.12.9/8.11.4) with SMTP id h6H6p1sL013773; Thu, 17 Jul 2003 09:51:05 +0300 (EEST) (envelope-from pete@he.iki.fi) Message-ID: <007501c34c2f$d20913d0$812a40c1@PETEX31> From: "Petri Helenius" To: "Jin Guojun [DSD]" References: <3F15C584.38848DCE@lbl.gov> <071401c34be3$71a03510$812a40c1@PETEX31> <00b501c34be5$351a83f0$b3db87d4@vader> <075101c34be5$db5e4c10$812a40c1@PETEX31> <3F15DE52.130D8FA9@lbl.gov> Date: Thu, 17 Jul 2003 09:50:55 +0300 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 8bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 cc: freebsd-performance@freebsd.org cc: Steven Hartland Subject: Re: timer counter chip access mystery] X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 17 Jul 2003 06:51:34 -0000 I thought you have to use ACPI-safe if you have broken hardware and othervise you´ll have ACPI-fast. (which my boxes default to) Pete ----- Original Message ----- From: "Jin Guojun [DSD]" To: "Petri Helenius" Cc: "Steven Hartland" ; Sent: Thursday, July 17, 2003 2:22 AM Subject: Re: timer counter chip access mystery] > 5.1 defaults to use ACPI-safe, which calls read_counter() at least > three times. The total cost is 3098 ns. So, it meas that read_count() > cost is about 1000 ns, which is close to what the Linux does. > > The rest question is how much error will be if not to use > ACPI-safe? i.e., sysctl kern.timecounter.hardware=ACPI? > > -Jin > > /* > * Fetch current time value from reliable hardware. > */ > static unsigned > acpi_timer_get_timecount(struct timecounter *tc) > { > return (read_counter()); > } > > /* > * Fetch current time value from hardware that may not correctly > * latch the counter. > */ > static unsigned > acpi_timer_get_timecount_safe(struct timecounter *tc) > { > unsigned u1, u2, u3; > > u2 = read_counter(); > u3 = read_counter(); > do { > u1 = u2; > u2 = u3; > u3 = read_counter(); > } while (u1 > u2 || u2 > u3 || (u3 - u1) > 15); > return (u2); > } > > Petri Helenius wrote: > > > 5.x defaults to ACPI or TSC timers. So your wish has been transported > > by a timemachine into the past. > > > > With 4.x you have to set it manually. > > > > Pete > > > > ----- Original Message ----- > > From: "Steven Hartland" > > To: "Petri Helenius" ; "Jin Guojun [DSD]" ; > > Sent: Thursday, July 17, 2003 12:57 AM > > Subject: Re: timer counter chip access mystery] > > > > > If this could be changed / fixed it would have a marked performance increase > > > for all game server code which use this function heavily from what I've seen > > > and would explain the sometimes huge performance differential between > > > FreeBSD and Linux. > > > > > > Steve / K > > > ----- Original Message ----- > > > From: "Petri Helenius" > > > To: "Jin Guojun [DSD]" ; > > > Sent: Wednesday, July 16, 2003 10:44 PM > > > Subject: Re: timer counter chip access mystery] > > > > > > > > > > > > > > This access happens over ISA bus and thus happens at the speed the bus > > > > operates at. Use TSC or ACPI for faster gettimeofday. > > > > > > > > Pete > > From owner-freebsd-performance@FreeBSD.ORG Thu Jul 17 03:49:45 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 860CA37B401 for ; Thu, 17 Jul 2003 03:49:45 -0700 (PDT) Received: from spf13.us4.outblaze.com (205-158-62-67.outblaze.com [205.158.62.67]) by mx1.FreeBSD.org (Postfix) with ESMTP id E10EE43F75 for ; Thu, 17 Jul 2003 03:49:44 -0700 (PDT) (envelope-from jwest254@mail.com) Received: from 205-158-62-68.outblaze.com (205-158-62-68.outblaze.com [205.158.62.68]) by spf13.us4.outblaze.com (Postfix) with QMQP id D2951181E419 for ; Thu, 17 Jul 2003 10:49:44 +0000 (GMT) Received: (qmail 71745 invoked from network); 17 Jul 2003 10:49:44 -0000 Received: from unknown (HELO ws1-8.us4.outblaze.com) (205.158.62.59) by 205-158-62-153.outblaze.com with SMTP; 17 Jul 2003 10:49:44 -0000 Received: (qmail 99328 invoked by uid 1001); 17 Jul 2003 10:49:44 -0000 Message-ID: <20030717104944.99327.qmail@mail.com> Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Content-Transfer-Encoding: 7bit MIME-Version: 1.0 X-Mailer: MIME-tools 5.41 (Entity 5.404) Received: from [193.120.103.96] by ws1-8.us4.outblaze.com with http for jwest254@mail.com; Thu, 17 Jul 2003 05:49:44 -0500 From: "James West" To: "Terry Lambert" Date: Thu, 17 Jul 2003 05:49:44 -0500 X-Originating-Ip: 193.120.103.96 X-Originating-Server: ws1-8.us4.outblaze.com cc: freebsd-performance@freebsd.org Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 17 Jul 2003 10:49:45 -0000 Added up all the sendq and recvq columns in a netstat -na output gives me almost always < 150 and i'm still getting almost constant no buffer space errors. The NIC was replaced only five minutes ago and the problem still persists. [root@flower] (/home/james): ping 127.0.0.1 ping: socket: No buffer space available [root@flower] (/home/james): ping 127.0.0.1 ping: socket: No buffer space available [root@flower] (/home/james): ping 127.0.0.1 ping: socket: No buffer space available [root@flower] (/home/james): ifconfig ifconfig: socket: No buffer space available [root@flower] (/home/james): ifconfig ifconfig: socket: No buffer space available [root@flower] (/home/james): dmesg | grep fxp0 fxp0: port 0xb000-0xb03f mem 0xda800000-0xda81ffff,0xdb000000-0xdb000fff irq 12 at device 16.0 on pci0 fxp0: Disabling dynamic standby mode in EEPROM fxp0: New EEPROM ID: 0x50a0 fxp0: EEPROM checksum @ 0x3f: 0xca9a -> 0xca9c fxp0: Ethernet address 00:02:b3:48:50:2a ----- Original Message ----- From: Terry Lambert Date: Wed, 16 Jul 2003 22:41:32 -0700 To: James West Subject: Re: Major problem with "No buffer space available" errors > James West wrote: > > I'm having huge problems with "No buffer space available" errors. I've > > increased MAXUSERS to 512 in the kernel, recompiled, rebooted and the > > sysctl values below show that everything is up'ed to the max. > > Historically, the most common cause of this message has been an > interface you were actively feeding data going down. Most commonly, > this occurs in the use of UDP and ICMP protocols (e.g. "ping") out a > downed interface, since they are not constrained by your sendspace > settings, which apply to TCP (they will basically permit you to use > all your buffer in the attempt). > > So the questions to answer are: > > 1) Are you using a lot of ICMP (e.g. ping, traceroute, RIP, etc.)? > > 2) Are you using a lot of UDP (e.g. Linux NFS clients using UDP > mounts and an rsize or wsize larger than the MTU would permit > to fit in a single UDP packet)? > > Other than that, you should "netstat -an" and add up the contents of > the SendQ/RecvQ columns. It's possible that, in fact, you *are* > running out of buffer space. For the default of 64K, you would need > 2G of RAM dedicated to nothing but mbufs (not including headers!) to > support only 32,768 simultaneous connections without mbuf overcommit. > > Dropping the sndspace/rcvspace paramaters back to their pre-bump-up > defaults will double the number of connections for the same amount > of RAM. > > -- Terry > _______________________________________________ > freebsd-performance@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-performance > To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" -- __________________________________________________________ Sign-up for your own FREE Personalized E-mail at Mail.com http://www.mail.com/?sr=signup CareerBuilder.com has over 400,000 jobs. Be smarter about your job search http://corp.mail.com/careers From owner-freebsd-performance@FreeBSD.ORG Thu Jul 17 03:52:54 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 565F037B401 for ; Thu, 17 Jul 2003 03:52:54 -0700 (PDT) Received: from spf13.us4.outblaze.com (205-158-62-67.outblaze.com [205.158.62.67]) by mx1.FreeBSD.org (Postfix) with ESMTP id 8888443FBF for ; Thu, 17 Jul 2003 03:52:53 -0700 (PDT) (envelope-from jwest254@mail.com) Received: from 205-158-62-68.outblaze.com (205-158-62-68.outblaze.com [205.158.62.68]) by spf13.us4.outblaze.com (Postfix) with QMQP id 7BE5F1821519 for ; Thu, 17 Jul 2003 10:52:53 +0000 (GMT) Received: (qmail 88174 invoked from network); 17 Jul 2003 10:52:53 -0000 Received: from unknown (HELO ws1-8.us4.outblaze.com) (205.158.62.59) by 205-158-62-153.outblaze.com with SMTP; 17 Jul 2003 10:52:53 -0000 Received: (qmail 3633 invoked by uid 1001); 17 Jul 2003 10:52:52 -0000 Message-ID: <20030717105252.3632.qmail@mail.com> Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Content-Transfer-Encoding: 7bit MIME-Version: 1.0 X-Mailer: MIME-tools 5.41 (Entity 5.404) Received: from [193.120.103.96] by ws1-8.us4.outblaze.com with http for jwest254@mail.com; Thu, 17 Jul 2003 05:52:52 -0500 From: "James West" To: "Terry Lambert" Date: Thu, 17 Jul 2003 05:52:52 -0500 X-Originating-Ip: 193.120.103.96 X-Originating-Server: ws1-8.us4.outblaze.com cc: freebsd-performance@freebsd.org Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 17 Jul 2003 10:52:54 -0000 > So the questions to answer are: > > 1) Are you using a lot of ICMP (e.g. ping, traceroute, RIP, etc.)? The box is almost idle - there's definitely no pings going on, and no traceroutes either, also no RIP. Could this be caused by a flood of some sort? (like a DoS attack?) > > 2) Are you using a lot of UDP (e.g. Linux NFS clients using UDP > mounts and an rsize or wsize larger than the MTU would permit > to fit in a single UDP packet)? No, not udp - the only UDP traffic that'd be going on with the box is DNS. The box runs qmail, apache, ssh and that's it. Nothing else - no other ports are open on the machine other than ones for these services, everything else is locked down (syslog's not listening on a network port, mysql isn't etc. etc.). -- __________________________________________________________ Sign-up for your own FREE Personalized E-mail at Mail.com http://www.mail.com/?sr=signup CareerBuilder.com has over 400,000 jobs. Be smarter about your job search http://corp.mail.com/careers From owner-freebsd-performance@FreeBSD.ORG Thu Jul 17 11:26:23 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 9B86237B401 for ; Thu, 17 Jul 2003 11:26:23 -0700 (PDT) Received: from salida.argentina.com (salida.argentina.com [64.76.16.54]) by mx1.FreeBSD.org (Postfix) with ESMTP id 3692443FB1 for ; Thu, 17 Jul 2003 11:26:22 -0700 (PDT) (envelope-from carlos@infodrive.com.ar) Received: from mail.infodrive.com.ar (unknown [200.68.192.253]) by salida.argentina.com (Postfix) with ESMTP id 4804C44C7DD for ; Thu, 17 Jul 2003 15:28:30 -0300 (ART) Received: from localhost (localhost.intradoc.com.ar [127.0.0.1]) by mail.infodrive.com.ar (Postfix) with ESMTP id AF33868DA8; Thu, 17 Jul 2003 15:26:32 -0300 (ART) Received: from 200.61.45.138 ( [200.61.45.138]) as user carlos@localhost by www.infodrive.com.ar with HTTP; Thu, 17 Jul 2003 15:26:32 -0300 Message-ID: <1058466392.3f16ea58729ba@www.infodrive.com.ar> Date: Thu, 17 Jul 2003 15:26:32 -0300 From: Carlos Horowicz To: James West References: <20030717105252.3632.qmail@mail.com> In-Reply-To: <20030717105252.3632.qmail@mail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit User-Agent: Internet Messaging Program (IMP) 3.1 / FreeBSD-4.7 X-Originating-IP: 200.61.45.138 X-MailScanner: Found to be clean, Found to be clean X-MailScanner-Information: Please contact the ISP for more information cc: freebsd-performance@freebsd.org cc: Terry Lambert Subject: Re: Major problem with "No buffer space available" errors X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 17 Jul 2003 18:26:24 -0000 Hi Mensaje citado por James West : > > > So the questions to answer are: > > > > 1) Are you using a lot of ICMP (e.g. ping, traceroute, RIP, etc.)? > > The box is almost idle - there's definitely no pings going on, and no > traceroutes either, also no RIP. > > Could this be caused by a flood of some sort? (like a DoS attack?) I had the same messages some time ago with a similar config : a 4.7 lan gateway doing nat, two NICs, one connected to an ADSL modem, the other one to the LAN ... it turned out to be a w2k in the LAN infected with sql-slammer, hitting my FreeBSD gateway, trying to connect to any ip address out in the internet scanning for MS-SQL servers... those bursts seemed to be eating up all network resources of the FreeBSD ... tcpdump helped diagnose the problem, ipfw solved it. Not sure if this has anything to do with your one NIC setup, but you might take a look at your traffic. Can't remember either whether the DoS was tcp, udp or both. > > > > > 2) Are you using a lot of UDP (e.g. Linux NFS clients using UDP > > mounts and an rsize or wsize larger than the MTU would permit > > to fit in a single UDP packet)? > > No, not udp - the only UDP traffic that'd be going on with the box is DNS. > The box runs qmail, apache, ssh and that's it. Nothing else - no other ports > are open on the machine other than ones for these services, everything else > is locked down (syslog's not listening on a network port, mysql isn't etc. > etc.). > > -- > __________________________________________________________ > Sign-up for your own FREE Personalized E-mail at Mail.com > http://www.mail.com/?sr=signup > > CareerBuilder.com has over 400,000 jobs. Be smarter about your job search > http://corp.mail.com/careers > > _______________________________________________ > freebsd-performance@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-performance > To unsubscribe, send any mail to > "freebsd-performance-unsubscribe@freebsd.org" > > -- > Este mensaje ha sido analizado en busca de virus y otros contenidos > peligrosos, y se considera que está limpio. > -- http://www.infodrive.com.ar - E-mail corporativo > -- Carlos ------------------------------------------------- Mail enviado con INFODRIVE -- Este mensaje ha sido analizado en busca de virus y otros contenidos peligrosos, y se considera que está limpio. -- http://www.infodrive.com.ar - E-mail corporativo -- Obtenga gratis su cuenta @Argentina.com con AntiVirus, 20mb de espacio y acceso libre UltraVeloz en 24 ciudades www.Argentina.com From owner-freebsd-performance@FreeBSD.ORG Thu Jul 17 21:00:45 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id AFC5837B404 for ; Thu, 17 Jul 2003 21:00:45 -0700 (PDT) Received: from web80006.mail.yahoo.com (web80006.mail.yahoo.com [66.163.168.136]) by mx1.FreeBSD.org (Postfix) with SMTP id D4A8843FA3 for ; Thu, 17 Jul 2003 21:00:44 -0700 (PDT) (envelope-from soundsampler@sbcglobal.net) Message-ID: <20030718040044.96606.qmail@web80006.mail.yahoo.com> Received: from [64.171.77.116] by web80006.mail.yahoo.com via HTTP; Thu, 17 Jul 2003 21:00:44 PDT Date: Thu, 17 Jul 2003 21:00:44 -0700 (PDT) From: Trent George To: sos@freebsd.org MIME-Version: 1.0 X-Mailman-Approved-At: Thu, 17 Jul 2003 21:13:58 -0700 Content-Type: text/plain; charset=us-ascii X-Content-Filtered-By: Mailman/MimeDel 2.1.1 cc: freebsd-performance@freebsd.org Subject: promise sx4000 card & tx4000 results over 170mb/s sustained read X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 18 Jul 2003 04:00:46 -0000 Hi Soren, I wanted to congratulate you on fantastic results in performance on a tx4000 card. I have a dual amd mb with 66mhz support slot. 4 western digital drives (180gb each) I have clocked sustained 176mb/s read from this raid0 array and 120+mb/s write writes seemed cpu bound ? (50% idle from second processor) reads 56% idle (first process close to maxed out) there were several tricks to getting this performance, custom kernel compile 1/ MAXPHYS=256k 2/ MAXBSIZE=256k 3/ atacontrol create RAID0 128 /dev/ad4 /dev/ad6 /dev/ad8 /dev/ad10 4/ newfs -U bs=262144 /dev/ar0 this allowed 64kb stripe size to use 4 drives on each "block" parallel reading i used iostat to monitor processor and transfer rate using following commands dd of=test.dat if=/dev/zero bs=64k count=20000 dd if=test.dat of=/dev/null bs=64k Question #2 I purchased a sx4000 card in the hope to increase speed :-) I noticed that PDC20621 is not in ata-chipset.c yet The card is not recognised on boot ether. Is there anything I can do or provide to help you add support for this card ? Question #3 Is there any plans to support intel SATA raid support on ICH5R (i865/i875) for boot up (sort of like promise and highpoint) I have been experimenting with gigabit ethernet and sustained 100mb/s samba read/write performance, and seem to be making good progress. also experimenting on huge writeback cache (400mb+) up from 1mb vfs.hirunningspace seems to help with impressive burst network transfer speed. Thanks Trent George PS I don't know if you remember but I sent an amd system to you a long time ago to help add support for highpoint chipsets, I appreciate all your great work. From owner-freebsd-performance@FreeBSD.ORG Fri Jul 18 01:15:53 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id F267F37B401; Fri, 18 Jul 2003 01:15:52 -0700 (PDT) Received: from spider.deepcore.dk (cpe.atm2-0-56339.0x50c6aa0a.abnxx2.customer.tele.dk [80.198.170.10]) by mx1.FreeBSD.org (Postfix) with ESMTP id AA10943F75; Fri, 18 Jul 2003 01:15:51 -0700 (PDT) (envelope-from sos@spider.deepcore.dk) Received: (from sos@localhost) by spider.deepcore.dk (8.12.9/8.12.9) id h6I8FnLd071918; Fri, 18 Jul 2003 10:15:49 +0200 (CEST) (envelope-from sos) From: Soeren Schmidt Message-Id: <200307180815.h6I8FnLd071918@spider.deepcore.dk> In-Reply-To: <20030718040044.96606.qmail@web80006.mail.yahoo.com> To: Trent George Date: Fri, 18 Jul 2003 10:15:49 +0200 (CEST) X-Mailer: ELM [version 2.4ME+ PL98b (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=ISO-8859-1 X-Mailman-Approved-At: Fri, 18 Jul 2003 04:51:06 -0700 cc: freebsd-performance@FreeBSD.ORG cc: sos@FreeBSD.ORG Subject: Re: promise sx4000 card & tx4000 results over 170mb/s sustained read X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 18 Jul 2003 08:15:53 -0000 It seems Trent George wrote: > Hi Soren, > > I wanted to congratulate you on fantastic results in performance on a tx4000 card. > > I have a dual amd mb with 66mhz support slot. 4 western digital drives (180gb each) > > I have clocked sustained 176mb/s read from this raid0 array and 120+mb/s write > > writes seemed cpu bound ? (50% idle from second processor) > reads 56% idle (first process close to maxed out) > > there were several tricks to getting this performance, custom kernel compile > > 1/ MAXPHYS=256k > 2/ MAXBSIZE=256k > 3/ atacontrol create RAID0 128 /dev/ad4 /dev/ad6 /dev/ad8 /dev/ad10 > 4/ newfs -U bs=262144 /dev/ar0 > > this allowed 64kb stripe size to use 4 drives on each "block" parallel reading > > i used iostat to monitor processor and transfer rate using following commands > > dd of=test.dat if=/dev/zero bs=64k count=20000 > dd if=test.dat of=/dev/null bs=64k Wooo! not bad at all :) I'd suspect performance like that, but dont have the HW to test it so this is most welcomed!! > Question #2 > > I purchased a sx4000 card in the hope to increase speed :-) > I noticed that PDC20621 is not in ata-chipset.c yet > The card is not recognised on boot ether. > > Is there anything I can do or provide to help you add support for this card ? I do have the sx4000 in the works (thanks to Promise who has provided docs and HW to make this happen).. > Question #3 > > Is there any plans to support intel SATA raid support on ICH5R (i865/i875) for boot up (sort of like promise and highpoint) Uhm, well, for this to happen I need to figure out the where they store the RAID config (my guess is on disk somewhere), then reverseengineer the contents and write routines for ata-raid.c to r/w the config blocks. This can be quite a bit of work and needs alot of trial error experimentation which needs someone with access to the HW and a fair amount of determination. (just FYI I've just ripped Adaptecs RAID config apart on their new SATA controller :) ) > PS I don't know if you remember but I sent an amd system to you a long time ago to help add support for highpoint chipsets, I appreciate all your great work. I do remember!! that system actually was my main workstation for a long time (until the capacitors gave up but that problem has hit alot of Abit boards from that time period), so it has contributed to FreeBSD in more than one sense, thanks!! -Søren From owner-freebsd-performance@FreeBSD.ORG Fri Jul 18 14:56:23 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 61E5837B401; Fri, 18 Jul 2003 14:56:23 -0700 (PDT) Received: from mta8.adelphia.net (mta8.adelphia.net [64.8.50.196]) by mx1.FreeBSD.org (Postfix) with ESMTP id 7DE4643F3F; Fri, 18 Jul 2003 14:56:22 -0700 (PDT) (envelope-from Support@Netflag.Net) Received: from nfn2.Netflag.Net ([68.69.240.35]) by mta8.adelphia.net (InterMail vM.5.01.05.32 201-253-122-126-132-20030307) with ESMTP id <20030718215621.YLWL27346.mta8.adelphia.net@nfn2.Netflag.Net>; Fri, 18 Jul 2003 17:56:21 -0400 Message-Id: <5.2.0.9.2.20030718145601.0243b7a0@pop.dc3.adelphia.net> X-Sender: pedramn@pop.dc3.adelphia.net X-Mailer: QUALCOMM Windows Eudora Version 5.2.0.9 Date: Fri, 18 Jul 2003 14:56:17 -0700 To: Trent George , sos@freebsd.org From: Pedram Nimreezi In-Reply-To: <20030718040044.96606.qmail@web80006.mail.yahoo.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed cc: freebsd-performance@freebsd.org Subject: Re: promise sx4000 card & tx4000 results over 170mb/s sustained read X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 18 Jul 2003 21:56:23 -0000 I have a raid10 configuration with 4 western digital 8mb cache 120gig hard drives using a 2400a controller from adaptec on a dual 2400 xeon with 2 gigs ECC ram I'm doing the same tests... I'm getting 33mb/s. I got the controller because I figured it would be better to use a hardware solution... Does all this sound right? I should be getting much better results shouldn't I? After striping and mirroing i get 220 gigs free and from a report I read this configuration gets better performance from the 2400A http://www.xbitlabs.com/articles/storage/display/adaptec-2400a.html At 09:00 PM 7/17/2003 -0700, Trent George wrote: >Hi Soren, > >I wanted to congratulate you on fantastic results in performance on a >tx4000 card. > >I have a dual amd mb with 66mhz support slot. 4 western digital drives >(180gb each) > >I have clocked sustained 176mb/s read from this raid0 array and 120+mb/s write > >writes seemed cpu bound ? (50% idle from second processor) >reads 56% idle (first process close to maxed out) > >there were several tricks to getting this performance, custom kernel compile > >1/ MAXPHYS=256k >2/ MAXBSIZE=256k >3/ atacontrol create RAID0 128 /dev/ad4 /dev/ad6 /dev/ad8 /dev/ad10 >4/ newfs -U bs=262144 /dev/ar0 > >this allowed 64kb stripe size to use 4 drives on each "block" parallel reading > >i used iostat to monitor processor and transfer rate using following commands > >dd of=test.dat if=/dev/zero bs=64k count=20000 >dd if=test.dat of=/dev/null bs=64k > >Question #2 > >I purchased a sx4000 card in the hope to increase speed :-) >I noticed that PDC20621 is not in ata-chipset.c yet >The card is not recognised on boot ether. > >Is there anything I can do or provide to help you add support for this card ? > >Question #3 > >Is there any plans to support intel SATA raid support on ICH5R (i865/i875) >for boot up (sort of like promise and highpoint) > >I have been experimenting with gigabit ethernet and sustained 100mb/s >samba read/write performance, and seem to be making good progress. > >also experimenting on huge writeback cache (400mb+) up from 1mb >vfs.hirunningspace >seems to help with impressive burst network transfer speed. > > >Thanks > >Trent George > > >PS I don't know if you remember but I sent an amd system to you a long >time ago to help add support for highpoint chipsets, I appreciate all your >great work. > >_______________________________________________ >freebsd-performance@freebsd.org mailing list >http://lists.freebsd.org/mailman/listinfo/freebsd-performance >To unsubscribe, send any mail to >"freebsd-performance-unsubscribe@freebsd.org"