From owner-freebsd-performance@FreeBSD.ORG Mon May 28 08:50:33 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 6E1A416A519 for ; Mon, 28 May 2007 08:50:33 +0000 (UTC) (envelope-from almarrie@gmail.com) Received: from an-out-0708.google.com (an-out-0708.google.com [209.85.132.250]) by mx1.freebsd.org (Postfix) with ESMTP id 2D98F13C4B0 for ; Mon, 28 May 2007 08:50:33 +0000 (UTC) (envelope-from almarrie@gmail.com) Received: by an-out-0708.google.com with SMTP id c14so451304anc for ; Mon, 28 May 2007 01:50:32 -0700 (PDT) DKIM-Signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:mime-version:content-type:content-transfer-encoding:content-disposition; b=dcY7qocGXv7Wb7qDPpFIBgxEkuJhhL3/cLLu8kEQiJHfSztgYWoila+v2yHDEQ5T/cC2kRVXgFKtJGV1lgXfzDVPy5Sz5E1DNEh8vOgXv3USSDtrf2rWmUOloay+RDHrGRNQDKkprDAkkiCRXQKWIKt25zSZIdrbicJIiY33Y4c= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:mime-version:content-type:content-transfer-encoding:content-disposition; b=jWjJGJvRH5B00Ky+YDkWy/kFcEsvxJ7zsawH+I9ECSvxRqa2umfoxqnR7EnCKczoAEVDxVwsa5VN1rJVi97HiUWDG5kqhfajClvH0qQoBUJKHYgZX3cg2Ow/9vOiY+zEq+Y5mQfUa1yn9nALsp7fAvono7X2Mm/ZL96I8vIrnCo= Received: by 10.100.202.13 with SMTP id z13mr4403945anf.1180340760687; Mon, 28 May 2007 01:26:00 -0700 (PDT) Received: by 10.100.9.14 with HTTP; Mon, 28 May 2007 01:26:00 -0700 (PDT) Message-ID: <499c70c0705280126o252d505q99ca5476bf4b6d62@mail.gmail.com> Date: Mon, 28 May 2007 11:26:00 +0300 From: "Abdullah Ibn Hamad Al-Marri" To: freebsd-performance@freebsd.org MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Subject: Tweaks tips for MySQL in FreeBSD 7.0-CURRENT SMP X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 28 May 2007 08:50:33 -0000 Hello Guys, I was able to use this command in 6.x Now when I tried it it doesn't work in 7.0 sysctl kern.threads.max_groups_per_proc=40000 What do you recommend for tweaking MySQL server 5.0.41 with MyISAM db in Dual Xeon 2.8 GHz server with 2 GB of ram and 10k rpm wcd raptor HD? -- -- Regards, -Abdullah Ibn Hamad Al-Marri Arab Portal http://www.WeArab.Net/ From owner-freebsd-performance@FreeBSD.ORG Wed May 30 18:25:12 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 440B916A4D7 for ; Wed, 30 May 2007 18:25:12 +0000 (UTC) (envelope-from gallatin@cs.duke.edu) Received: from duke.cs.duke.edu (duke.cs.duke.edu [152.3.140.1]) by mx1.freebsd.org (Postfix) with ESMTP id 09BF313C469 for ; Wed, 30 May 2007 18:25:11 +0000 (UTC) (envelope-from gallatin@cs.duke.edu) Received: from grasshopper.cs.duke.edu (grasshopper.cs.duke.edu [152.3.145.30]) by duke.cs.duke.edu (8.14.0/8.14.0) with ESMTP id l4UIPBqG006966 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Wed, 30 May 2007 14:25:11 -0400 (EDT) Received: (from gallatin@localhost) by grasshopper.cs.duke.edu (8.12.9p2/8.12.9/Submit) id l4UIOhHY036660; Wed, 30 May 2007 14:24:43 -0400 (EDT) (envelope-from gallatin) From: Andrew Gallatin MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <18013.49538.356407.283631@grasshopper.cs.duke.edu> Date: Wed, 30 May 2007 14:24:43 -0400 (EDT) To: security In-Reply-To: <4652383E.9000302@jim-liesl.org> References: <4652383E.9000302@jim-liesl.org> X-Mailer: VM 6.75 under 21.1 (patch 12) "Channel Islands" XEmacs Lucid Cc: freebsd-performance@freebsd.org Subject: Re: [Fwd: asymetric speeds over gigE link] X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 30 May 2007 18:25:12 -0000 security writes: > Sent this to -net and didn't get much info, so I'll try here since > there's some overlap > > Summary: Using iperf to measure TCP net speed between a linux (kubuntu edgy) and Iperf is probably your problem, it tends to perform really poorly on nearly any OS other than linux as it measures lots of things besides network performance (like gettimeofday performance). Try a network benchmark that just benchmarks the network, like netperf. For example, on an amd64x2 running -current, I see ~4.75Gb/s from netperf (no options) and 2.85Gb/s from iperf (no options) Drew % netperf242 -H127.0.0.1 TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 127.0.0.1 (127.0.0.1) port 0 AF_INET Recv Send Send Socket Socket Message Elapsed Size Size Size Time Throughput bytes bytes bytes secs. 10^6bits/sec 65536 32768 32768 10.00 4757.03 % iperf -c 127.0.0.1 ------------------------------------------------------------ Client connecting to 127.0.0.1, TCP port 5001 TCP window size: 42.0 KByte (default) ------------------------------------------------------------ [ 3] local 127.0.0.1 port 60961 connected with 127.0.0.1 port 5001 [ 3] 0.0-10.0 sec 3.31 GBytes 2.85 Gbits/sec From owner-freebsd-performance@FreeBSD.ORG Wed May 30 20:49:49 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 92D4316A468 for ; Wed, 30 May 2007 20:49:49 +0000 (UTC) (envelope-from security@jim-liesl.org) Received: from qsmtp3.mc.surewest.net (qsmtp.mc.surewest.net [66.60.130.145]) by mx1.freebsd.org (Postfix) with SMTP id 7579313C447 for ; Wed, 30 May 2007 20:49:49 +0000 (UTC) (envelope-from security@jim-liesl.org) Received: (qmail 8964 invoked from network); 30 May 2007 13:49:48 -0700 Received: by simscan 1.1.0 ppid: 8929, pid: 8933, t: 4.5192s scanners: regex: 1.1.0 attach: 1.1.0 clamav: 0.84/m:43/d:3122 spam: 3.0.3 Received: from unknown (HELO daemon.jim-liesl.org) (66.60.173.44) by qsmtp3 with SMTP; 30 May 2007 13:49:44 -0700 Received: from daemon.jim-liesl.org (localhost.static.surewest.net [127.0.0.1]) by daemon.jim-liesl.org (Postfix) with ESMTP id 56BC65C1D; Wed, 30 May 2007 13:49:43 -0700 (PDT) Received: from [127.0.0.1] (daemon.static.surewest.net [192.168.1.15]) by daemon.jim-liesl.org (Postfix) with ESMTP id 122595C1C; Wed, 30 May 2007 13:49:43 -0700 (PDT) Message-ID: <465DE367.8070009@jim-liesl.org> Date: Wed, 30 May 2007 13:49:43 -0700 From: security User-Agent: Thunderbird 1.5.0.10 (Windows/20070221) MIME-Version: 1.0 To: Andrew Gallatin References: <4652383E.9000302@jim-liesl.org> <18013.49538.356407.283631@grasshopper.cs.duke.edu> In-Reply-To: <18013.49538.356407.283631@grasshopper.cs.duke.edu> X-Enigmail-Version: 0.94.1.1 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit X-Virus-Scanned: ClamAV using ClamSMTP X-Spam-Checker-Version: SpamAssassin 3.0.3 (2005-04-27) on qsmtp3.surewest.net X-Spam-Level: X-Spam-Status: No, score=-2.3 required=5.0 tests=AWL,BAYES_00 autolearn=ham version=3.0.3 Cc: freebsd-performance@freebsd.org Subject: Re: [Fwd: asymetric speeds over gigE link] X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 30 May 2007 20:49:49 -0000 > security writes: > > Sent this to -net and didn't get much info, so I'll try here since > > there's some overlap > > > > Summary: Using iperf to measure TCP net speed between a linux (kubuntu edgy) and > > Iperf is probably your problem, it tends to perform really poorly on > nearly any OS other than linux as it measures lots of things besides > network performance (like gettimeofday performance). Try a network > benchmark that just benchmarks the network, like netperf. > > For example, on an amd64x2 running -current, I see ~4.75Gb/s from > netperf (no options) and 2.85Gb/s from iperf (no options) > > Drew > Drew, Odd you should mention netperf. It yielded nearly identical asymmetric results as iperf. I ended up testing with netpipe (NPtcp). It reported fairly symmetric speeds between the two boxes when the send and recv buffers were set the same, so that mystery is solved (maybe). That being said I still wonder if I've too abstracted the net perfromance. The down side is that the best speed I got was around 362Mbps on a gigE link (with netpipe). I had hoped for better. From owner-freebsd-performance@FreeBSD.ORG Thu May 31 02:44:45 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 8A8C116A400 for ; Thu, 31 May 2007 02:44:45 +0000 (UTC) (envelope-from gnn@neville-neil.com) Received: from mrout1.yahoo.com (mrout1.yahoo.com [216.145.54.171]) by mx1.freebsd.org (Postfix) with ESMTP id 7748413C45E for ; Thu, 31 May 2007 02:44:45 +0000 (UTC) (envelope-from gnn@neville-neil.com) Received: from minion.local.neville-neil.com (proxy8.corp.yahoo.com [216.145.48.13]) by mrout1.yahoo.com (8.13.6/8.13.6/y.out) with ESMTP id l4V2YNSh000898; Wed, 30 May 2007 19:34:24 -0700 (PDT) Date: Thu, 31 May 2007 11:09:18 +0900 Message-ID: From: "George V. Neville-Neil" To: security In-Reply-To: <465DE367.8070009@jim-liesl.org> References: <4652383E.9000302@jim-liesl.org> <18013.49538.356407.283631@grasshopper.cs.duke.edu> <465DE367.8070009@jim-liesl.org> User-Agent: Wanderlust/2.14.0 (Africa) SEMI/1.14.6 (Maruoka) FLIM/1.14.8 (=?ISO-8859-4?Q?Shij=F2?=) APEL/10.7 Emacs/22.0.95 (i386-apple-darwin8.8.2) MULE/5.0 (SAKAKI) MIME-Version: 1.0 (generated by SEMI 1.14.6 - "Maruoka") Content-Type: text/plain; charset=US-ASCII X-Mailman-Approved-At: Thu, 31 May 2007 04:30:23 +0000 Cc: freebsd-performance@freebsd.org, Andrew Gallatin Subject: Re: [Fwd: asymetric speeds over gigE link] X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 31 May 2007 02:44:45 -0000 At Wed, 30 May 2007 13:49:43 -0700, security wrote: > Drew, > Odd you should mention netperf. It yielded nearly identical > asymmetric results as iperf. I ended up testing with netpipe (NPtcp). > It reported fairly symmetric speeds between the two boxes when the send > and recv buffers were set the same, so that mystery is solved (maybe). > That being said I still wonder if I've too abstracted the net perfromance. > > The down side is that the best speed I got was around 362Mbps on > a gigE link (with netpipe). I had hoped for better. NetPIPE is my test of choice for such things, but I have not fully evaluated where it's performance bottlenecks may lie. It is most useful for causing "problems" to be found with networking code because unlike most tests it attempts to use odd sized packets. An interesting experiment would be to run gprof on NetPIPE to make sure that it was not the source of the inefficiency. If you try that can you post results? I'm the maintainer of the NetPIPE port in FreeBSD and have put in the code to cover IPv6 and SCTP. Best, George From owner-freebsd-performance@FreeBSD.ORG Thu May 31 10:26:27 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 1A3F316A400 for ; Thu, 31 May 2007 10:26:27 +0000 (UTC) (envelope-from gallatin@cs.duke.edu) Received: from duke.cs.duke.edu (duke.cs.duke.edu [152.3.140.1]) by mx1.freebsd.org (Postfix) with ESMTP id BD0ED13C44B for ; Thu, 31 May 2007 10:26:26 +0000 (UTC) (envelope-from gallatin@cs.duke.edu) Received: from grasshopper.cs.duke.edu (grasshopper.cs.duke.edu [152.3.145.30]) by duke.cs.duke.edu (8.14.0/8.14.0) with ESMTP id l4VAQQd0009372 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Thu, 31 May 2007 06:26:26 -0400 (EDT) Received: (from gallatin@localhost) by grasshopper.cs.duke.edu (8.12.9p2/8.12.9/Submit) id l4VAPvtr037698; Thu, 31 May 2007 06:25:57 -0400 (EDT) (envelope-from gallatin) From: Andrew Gallatin MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <18014.41676.681948.864061@grasshopper.cs.duke.edu> Date: Thu, 31 May 2007 06:25:57 -0400 (EDT) To: security In-Reply-To: <465DE367.8070009@jim-liesl.org> References: <4652383E.9000302@jim-liesl.org> <18013.49538.356407.283631@grasshopper.cs.duke.edu> <465DE367.8070009@jim-liesl.org> X-Mailer: VM 6.75 under 21.1 (patch 12) "Channel Islands" XEmacs Lucid Cc: freebsd-performance@freebsd.org Subject: Re: [Fwd: asymetric speeds over gigE link] X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 31 May 2007 10:26:27 -0000 security writes: > The down side is that the best speed I got was around 362Mbps on a > gigE link (with netpipe). I had hoped for better. 362Mb/s is way too low for 1GbE. You might want to try a UDP test to isolate the bottleneck somewhat. Try netperf -tUDP_STREAM -Hmacpro02-m -- -m 1472 and see what is reported for the transmit vs receive throughput. Drew From owner-freebsd-performance@FreeBSD.ORG Fri Jun 1 10:38:44 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 4A8A716A400 for ; Fri, 1 Jun 2007 10:38:44 +0000 (UTC) (envelope-from hunreal@gmail.com) Received: from an-out-0708.google.com (an-out-0708.google.com [209.85.132.241]) by mx1.freebsd.org (Postfix) with ESMTP id 0D00413C43E for ; Fri, 1 Jun 2007 10:38:43 +0000 (UTC) (envelope-from hunreal@gmail.com) Received: by an-out-0708.google.com with SMTP id c14so194537anc for ; Fri, 01 Jun 2007 03:38:43 -0700 (PDT) DKIM-Signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:mime-version:content-type:content-transfer-encoding:content-disposition; b=QQLlUv3OEt0KuTYWq693YIjLw8CiX+wbsK0ERzI3gM0+1Lj5Nanl1EDwzI3nnFD0UbTpsOhyVRXrRj27iE70LVs45nTT9939DjtrHIVAS4oXVpC/0g5bwTgDWwYRuaCLopwZylQAjglzZoJ383TbV43GtmgyPTUM+7pNHnooz4c= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:mime-version:content-type:content-transfer-encoding:content-disposition; b=byYf7QQ0DffxXURQyjETBcumkQBfXGkUNWO6SRFFrrxOFy/CuN9IiH4sOnbY+d9QcMrPSnbKP0ftl+aWsLgkdEIbPZWBPxhxr3NuLYPAiH1bu+eKe17q/PgvGaYa2wbKjtO019yvEYKC78V3xyABULvTr9tG8YFgeDjJHKyZfv0= Received: by 10.100.40.17 with SMTP id n17mr938839ann.1180692653164; Fri, 01 Jun 2007 03:10:53 -0700 (PDT) Received: by 10.100.95.14 with HTTP; Fri, 1 Jun 2007 03:10:53 -0700 (PDT) Message-ID: <9b6b59500706010310h145a1804ie97e57aa0c5e90a4@mail.gmail.com> Date: Fri, 1 Jun 2007 18:10:53 +0800 From: hshh To: freebsd-performance@freebsd.org MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Subject: Bad performance while transfer large block size through NFS. X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 01 Jun 2007 10:38:44 -0000 I set up one NFS server, and mounted on other server by TCP. Servers connected with Giga network, and running 6.2-RELEASE. But I found the performance is very bad while transfering large block size data. For example, I use dd on NFS client to test the speed. # dd if=/dev/zero of=/mnt/test bs=16k count=20k 20480+0 records in 20480+0 records out 335544320 bytes transferred in 6.172289 secs (54363027 bytes/sec) # dd if=/dev/zero of=/mnt/test bs=32k count=10k 10240+0 records in 10240+0 records out 335544320 bytes transferred in 6.481602 secs (51768733 bytes/sec) # dd if=/dev/zero of=/mnt/test bs=512k count=1k 1024+0 records in 1024+0 records out 536870912 bytes transferred in 11.859133 secs (45270671 bytes/sec) # dd if=/dev/zero of=/mnt/test bs=1m count=512 512+0 records in 512+0 records out 536870912 bytes transferred in 11.963019 secs (44877544 bytes/sec) # dd if=/dev/zero of=/mnt/test bs=1651k count=200 200+0 records in 200+0 records out 338124800 bytes transferred in 7.431332 secs (45499891 bytes/sec) # dd if=/dev/zero of=/mnt/test bs=1652k count=50 50+0 records in 50+0 records out 84582400 bytes transferred in 22.319528 secs (3789614 bytes/sec) ----Test End--- As test result, while block size smaller than 1652k, it's fast. But speed drop to only 3.6M/s while block size bigger or equal 1652k. And ``systat -vm 1'' is displayed HDD is most busy, Disks da0 KB/t 16.00 tps 239 MB/s 3.73 % busy 91 Can I do some tuning to improve large block size transfer over NFS? From owner-freebsd-performance@FreeBSD.ORG Fri Jun 1 11:27:25 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id E011B16A400 for ; Fri, 1 Jun 2007 11:27:25 +0000 (UTC) (envelope-from kometen@gmail.com) Received: from wa-out-1112.google.com (wa-out-1112.google.com [209.85.146.181]) by mx1.freebsd.org (Postfix) with ESMTP id B968A13C480 for ; Fri, 1 Jun 2007 11:27:25 +0000 (UTC) (envelope-from kometen@gmail.com) Received: by wa-out-1112.google.com with SMTP id m33so638193wag for ; Fri, 01 Jun 2007 04:27:25 -0700 (PDT) DKIM-Signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=DspBMLsnHMoTLf8SyONWcABWyf2oUUAMWH5uilLqhFKFvFO+lK7gYxVTeCczUwvU30vZ2H7XmBp/e7u+nNcCpIK6icdNjlLT+O8iek9G/UHTf9urq9Eby5VZss1w9jmK4Pvec/6xUeYxdnGpNII0uxfuMahXQ5IwxB7xXXmRRp0= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=GmLFc7hsyifN2DTr3LKDHYKyEPrtvmxRHodGKFP+AoUCSe99GJUoJAmFxmjD/9qNnMixnXJ9rca73zSyJpIH+nE91H9mfVgynuWPjldQxP8VxG6Ud95P+7ysfHUqw6WDP8oN3kn0xarhAsELgitYCIQjWwEasEDmaX6Fk9sKeKA= Received: by 10.114.201.1 with SMTP id y1mr1634895waf.1180695791266; Fri, 01 Jun 2007 04:03:11 -0700 (PDT) Received: by 10.114.194.12 with HTTP; Fri, 1 Jun 2007 04:03:11 -0700 (PDT) Message-ID: Date: Fri, 1 Jun 2007 13:03:11 +0200 From: "Claus Guttesen" To: hshh In-Reply-To: <9b6b59500706010310h145a1804ie97e57aa0c5e90a4@mail.gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <9b6b59500706010310h145a1804ie97e57aa0c5e90a4@mail.gmail.com> Cc: freebsd-performance@freebsd.org Subject: Re: Bad performance while transfer large block size through NFS. X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 01 Jun 2007 11:27:26 -0000 > I set up one NFS server, and mounted on other server by TCP. Servers > connected with Giga network, and running 6.2-RELEASE. > > But I found the performance is very bad while transfering large block size data. > For example, I use dd on NFS client to test the speed. > And ``systat -vm 1'' is displayed HDD is most busy, > Disks da0 > KB/t 16.00 > tps 239 > MB/s 3.73 > % busy 91 > > Can I do some tuning to improve large block size transfer over NFS? Try adding read- and write-size 32768, partial /etc/fstab: rw,intr,nfsv3,-a=4,-w=32768,-r=32768 or adding -r 32768 -w 32768 while mounting from command line. -- regards Claus When lenity and cruelty play for a kingdom, the gentlest gamester is the soonest winner. Shakespeare From owner-freebsd-performance@FreeBSD.ORG Fri Jun 1 13:11:36 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id BF3E916A400 for ; Fri, 1 Jun 2007 13:11:36 +0000 (UTC) (envelope-from tec@mega.net.br) Received: from msrv.matik.com.br (msrv.matik.com.br [200.152.83.14]) by mx1.freebsd.org (Postfix) with ESMTP id 3D04E13C44C for ; Fri, 1 Jun 2007 13:11:35 +0000 (UTC) (envelope-from tec@mega.net.br) Received: from ap-h.matik.com.br (ap-h.matik.com.br [200.152.83.36]) by msrv.matik.com.br (8.14.1/8.13.1) with ESMTP id l51BuXKu076726 for ; Fri, 1 Jun 2007 08:56:33 -0300 (BRT) (envelope-from tec@mega.net.br) From: NOC Meganet Organization: Prowip Telecom Ltda To: freebsd-performance@freebsd.org Date: Fri, 1 Jun 2007 08:55:59 -0300 User-Agent: KMail/1.9.6 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200706010855.59841.tec@mega.net.br> X-Spam-Status: No, score=-100.2 required=5.0 tests=ALL_TRUSTED, AWL, ISO_7BITS, USER_IN_WHITELIST autolearn=unavailable version=3.1.8 X-Spam-Checker-Version: Antispam Datacenter Matik msrv.matik.com.br X-Virus-Scanned: ClamAV version 0.88.4, clamav-milter version 0.88.4 on msrv.matik.com.br X-Virus-Status: Clean Subject: hd device names X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 01 Jun 2007 13:11:36 -0000 Hi all I am not sure if this is exactly the right list but perhaps some could point me to it. I think it is not the best way having different hard disk device names. I think it should be generally hd or something independent of type, driver or controller. Once I had trouble because I needed to change the MB and the new one called the first SATA ad8 and not ad4 Now I was terrible bitten by a ahd controller failure and I had no suitable backup card on site and decided to put an aac card in, so here it was not da anymore it was aacd which caused me a long down time of my cluster because the hardware maintainer was not trained to get the fstab change done. So long story short question, is there any link trick so that I can prepare my servers for something like that otherwise I would like to make such a suggestion to the project people. Meyer -- -- Prowip Telecom Ltda AS 22706 A mensagem foi scaneada pelo sistema de e-mail e pode ser considerada segura. Service fornecido pelo Datacenter Matik https://datacenter.matik.com.br From owner-freebsd-performance@FreeBSD.ORG Fri Jun 1 13:41:06 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 0DEB916A41F for ; Fri, 1 Jun 2007 13:41:06 +0000 (UTC) (envelope-from arne_woerner@yahoo.com) Received: from web30302.mail.mud.yahoo.com (web30302.mail.mud.yahoo.com [209.191.69.64]) by mx1.freebsd.org (Postfix) with SMTP id B036F13C457 for ; Fri, 1 Jun 2007 13:41:05 +0000 (UTC) (envelope-from arne_woerner@yahoo.com) Received: (qmail 14781 invoked by uid 60001); 1 Jun 2007 13:14:24 -0000 DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=X-YMail-OSG:Received:Date:From:Subject:To:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding:Message-ID; b=h4xo+H0sL+hI2Hht/rGqZYnVlaQT/KPXyntoU9W1g7kucJlufjJOucPVufKX32yXKydcM2fYKJbIJUt2zC3FTydfAPMK/o+LAmY/Nc/fjTaO2uE0I69IYo0TrlnnIPAR3khYq9cV/tWtv+MQi0VBPhHRcNUw0aRi+EcpfhhAKeQ=; X-YMail-OSG: B.XiqkIVM1lMRym.5akf3eZm3VJOx96wA27jxBL. Received: from [77.129.186.25] by web30302.mail.mud.yahoo.com via HTTP; Fri, 01 Jun 2007 06:14:23 PDT Date: Fri, 1 Jun 2007 06:14:23 -0700 (PDT) From: Arne "Wörner" To: NOC Meganet , freebsd-performance@freebsd.org In-Reply-To: <200706010855.59841.tec@mega.net.br> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Message-ID: <908099.12553.qm@web30302.mail.mud.yahoo.com> Cc: Subject: Re: hd device names X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 01 Jun 2007 13:41:06 -0000 --- NOC Meganet wrote: > I think it is not the best way having different hard disk device names. I > think it should be generally hd or something independent of type, driver or > controller. > Why dönt u use glabel? -Arne ____________________________________________________________________________________ Luggage? GPS? Comic books? Check out fitting gifts for grads at Yahoo! Search http://search.yahoo.com/search?fr=oni_on_mail&p=graduation+gifts&cs=bz From owner-freebsd-performance@FreeBSD.ORG Fri Jun 1 13:55:36 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 283FA16A400 for ; Fri, 1 Jun 2007 13:55:36 +0000 (UTC) (envelope-from tec@mega.net.br) Received: from msrv.matik.com.br (msrv.matik.com.br [200.152.83.14]) by mx1.freebsd.org (Postfix) with ESMTP id A32DD13C46A for ; Fri, 1 Jun 2007 13:55:35 +0000 (UTC) (envelope-from tec@mega.net.br) Received: from anb.matik.com.br (anb.matik.com.br [200.152.83.34]) by msrv.matik.com.br (8.14.1/8.13.1) with ESMTP id l51DtUFd089834; Fri, 1 Jun 2007 10:55:30 -0300 (BRT) (envelope-from tec@mega.net.br) From: NOC Prowip Organization: Prowip Telecom Ltda To: freebsd-performance@freebsd.org Date: Fri, 1 Jun 2007 10:55:22 -0300 User-Agent: KMail/1.9.6 References: <908099.12553.qm@web30302.mail.mud.yahoo.com> In-Reply-To: <908099.12553.qm@web30302.mail.mud.yahoo.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Message-Id: <200706011055.24039.tec@mega.net.br> X-Spam-Status: No, score=-100.0 required=5.0 tests=ALL_TRUSTED, AWL, MR_DIFF_MID, USER_IN_WHITELIST autolearn=unavailable version=3.1.8 X-Spam-Checker-Version: Antispam Datacenter Matik msrv.matik.com.br X-Virus-Scanned: ClamAV version 0.88.4, clamav-milter version 0.88.4 on msrv.matik.com.br X-Virus-Status: Clean Cc: Arne =?iso-8859-1?q?W=F6rner?= Subject: Re: hd device names X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 01 Jun 2007 13:55:36 -0000 On Friday 01 June 2007 10:14:23 Arne W=F6rner wrote: > --- NOC Meganet wrote: > > I think it is not the best way having different hard disk device names.= I > > think it should be generally hd or something independent of type, driver > > or controller. > > Why d=F6nt u use glabel? > hmm, I never considered it but I guess it does not solve my problem because let's say da0 is the=20 provider and da0 might have still the label I gave it but the OS can not se= e=20 it since it is then aacd0 after changing the controller or am I wrong? Meyer =2D-=20 =2D- Prowip Telecom Ltda AS 22706 A mensagem foi scaneada pelo sistema de e-mail e pode ser considerada segura. Service fornecido pelo Datacenter Matik https://datacenter.matik.com.br From owner-freebsd-performance@FreeBSD.ORG Fri Jun 1 14:30:25 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 5765616A400 for ; Fri, 1 Jun 2007 14:30:25 +0000 (UTC) (envelope-from arne_woerner@yahoo.com) Received: from web30307.mail.mud.yahoo.com (web30307.mail.mud.yahoo.com [209.191.69.69]) by mx1.freebsd.org (Postfix) with SMTP id 09F3513C48A for ; Fri, 1 Jun 2007 14:30:24 +0000 (UTC) (envelope-from arne_woerner@yahoo.com) Received: (qmail 64494 invoked by uid 60001); 1 Jun 2007 14:30:24 -0000 DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=X-YMail-OSG:Received:Date:From:Subject:To:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding:Message-ID; b=qVaZhno65t7irY7GPChUpp3mhlZ+Q9radYdSd3UN3PMgZr42xvENGGQuhMOn+MUk24o7BBhFYBg4OpQqWzT+aKNDhy7MobKRFEbOdxiWIR5DKif8qJuN4xnxw5fYtPPc2Pqeu/Id7++UMylzU3PHmDDQ6IAS7qFLddaEcMbY4Uk=; X-YMail-OSG: zgDaVN0VM1kXzxBxHxe4oJCF61E6pcXKCm3QdztRAZNVBWhiBzAPUHvhEHEuAECMKHx9mGdeIQ.p8Vj6VUUA.m0T6KZIaGBwgWa6MRQRCxXsmbiRa2VEj8D9e3AZsw-- Received: from [77.129.186.25] by web30307.mail.mud.yahoo.com via HTTP; Fri, 01 Jun 2007 07:30:24 PDT Date: Fri, 1 Jun 2007 07:30:24 -0700 (PDT) From: Arne "Wörner" To: NOC Prowip , freebsd-performance@freebsd.org In-Reply-To: <200706011055.24039.tec@mega.net.br> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Message-ID: <506722.42229.qm@web30307.mail.mud.yahoo.com> Cc: Subject: Re: hd device names X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 01 Jun 2007 14:30:25 -0000 --- NOC Prowip wrote: > > Why dönt u use glabel? > > > hmm, I never considered it > but I guess it does not solve my problem because let's say da0 is the > provider and da0 might have still the label I gave it but the OS can not see > it since it is then aacd0 after changing the controller or am I wrong? > Hmm... Advisory: :-)) 1. Just try it. Dont believe me without thorough testing... :-) E. g. with (a) mdconfig -a -u 0 -f testdevfile ; glabel label -v fook md0 ; mdconfig -d -u 0 ; mdconfig -a -u 69 -f testdevfile ; ls /dev/label 2. I think that is one advantage of glabel: It does not care for the device name, but it just tastes the last sector of each device, as soon as it shows up (e. g. after the reboot). -Arne ____________________________________________________________________________________ Get the Yahoo! toolbar and be alerted to new email wherever you're surfing. http://new.toolbar.yahoo.com/toolbar/features/mail/index.php From owner-freebsd-performance@FreeBSD.ORG Fri Jun 1 19:46:22 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id E0D1816A468 for ; Fri, 1 Jun 2007 19:46:22 +0000 (UTC) (envelope-from tec@mega.net.br) Received: from msrv.matik.com.br (msrv.matik.com.br [200.152.83.14]) by mx1.freebsd.org (Postfix) with ESMTP id 50D1313C46C for ; Fri, 1 Jun 2007 19:46:22 +0000 (UTC) (envelope-from tec@mega.net.br) Received: from ap-h.matik.com.br (ap-h.matik.com.br [200.152.83.36]) by msrv.matik.com.br (8.14.1/8.13.1) with ESMTP id l51JkKPh024335; Fri, 1 Jun 2007 16:46:20 -0300 (BRT) (envelope-from tec@mega.net.br) From: NOC Meganet Organization: Prowip Telecom Ltda To: freebsd-performance@freebsd.org Date: Fri, 1 Jun 2007 16:45:47 -0300 User-Agent: KMail/1.9.6 References: <506722.42229.qm@web30307.mail.mud.yahoo.com> In-Reply-To: <506722.42229.qm@web30307.mail.mud.yahoo.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Message-Id: <200706011645.48111.tec@mega.net.br> X-Spam-Status: No, score=-100.2 required=5.0 tests=ALL_TRUSTED, AWL, MR_DIFF_MID, SMILEY, TW_EV, TW_VF, USER_IN_WHITELIST autolearn=unavailable version=3.1.8 X-Spam-Checker-Version: Antispam Datacenter Matik msrv.matik.com.br X-Virus-Scanned: ClamAV version 0.88.4, clamav-milter version 0.88.4 on msrv.matik.com.br X-Virus-Status: Clean Cc: Arne =?iso-8859-1?q?W=F6rner?= Subject: Re: hd device names X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 01 Jun 2007 19:46:23 -0000 On Friday 01 June 2007 11:30:24 Arne W=F6rner wrote: > --- NOC Prowip wrote: > > > Why d=F6nt u use glabel? > > > > hmm, I never considered it > > but I guess it does not solve my problem because let's say da0 is the > > provider and da0 might have still the label I gave it but the OS can not > > see it since it is then aacd0 after changing the controller or am I > > wrong? > > Hmm... > > Advisory: :-)) > 1. > Just try it. Dont believe me without thorough testing... :-) > E. g. with (a) mdconfig -a -u 0 -f testdevfile ; glabel label -v fook md0= ; > mdconfig -d -u 0 ; mdconfig -a -u 69 -f testdevfile ; ls /dev/label > > 2. > I think that is one advantage of glabel: It does not care for the device > name, but it just tastes the last sector of each device, as soon as it > shows up (e. g. after the reboot). > thank's Arne, I will give it a try on a test machine and give you a feedbac= k=20 on it =2D- Prowip Telecom Ltda AS 22706 A mensagem foi scaneada pelo sistema de e-mail e pode ser considerada segura. Service fornecido pelo Datacenter Matik https://datacenter.matik.com.br From owner-freebsd-performance@FreeBSD.ORG Fri Jun 1 19:53:20 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id F368316A488 for ; Fri, 1 Jun 2007 19:53:19 +0000 (UTC) (envelope-from kris@obsecurity.org) Received: from elvis.mu.org (elvis.mu.org [192.203.228.196]) by mx1.freebsd.org (Postfix) with ESMTP id C5FB513C4C4 for ; Fri, 1 Jun 2007 19:53:19 +0000 (UTC) (envelope-from kris@obsecurity.org) Received: from obsecurity.dyndns.org (elvis.mu.org [192.203.228.196]) by elvis.mu.org (Postfix) with ESMTP id 3F2A81A3C1A; Fri, 1 Jun 2007 12:54:32 -0700 (PDT) Received: from rot13.obsecurity.org (rot13.obsecurity.org [192.168.1.5]) by obsecurity.dyndns.org (Postfix) with ESMTP id 10264511A7; Fri, 1 Jun 2007 15:53:19 -0400 (EDT) Received: by rot13.obsecurity.org (Postfix, from userid 1001) id F2247C1D6; Fri, 1 Jun 2007 15:53:18 -0400 (EDT) Date: Fri, 1 Jun 2007 15:53:18 -0400 From: Kris Kennaway To: NOC Meganet Message-ID: <20070601195318.GA57337@rot13.obsecurity.org> References: <506722.42229.qm@web30307.mail.mud.yahoo.com> <200706011645.48111.tec@mega.net.br> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200706011645.48111.tec@mega.net.br> User-Agent: Mutt/1.4.2.2i Cc: freebsd-performance@freebsd.org, Arne W?rner Subject: Re: hd device names X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 01 Jun 2007 19:53:20 -0000 On Fri, Jun 01, 2007 at 04:45:47PM -0300, NOC Meganet wrote: > On Friday 01 June 2007 11:30:24 Arne W?rner wrote: > > --- NOC Prowip wrote: > > > > Why d?nt u use glabel? > > > > > > hmm, I never considered it > > > but I guess it does not solve my problem because let's say da0 is the > > > provider and da0 might have still the label I gave it but the OS can not > > > see it since it is then aacd0 after changing the controller or am I > > > wrong? > > > > Hmm... > > > > Advisory: :-)) > > 1. > > Just try it. Dont believe me without thorough testing... :-) > > E. g. with (a) mdconfig -a -u 0 -f testdevfile ; glabel label -v fook md0 ; > > mdconfig -d -u 0 ; mdconfig -a -u 69 -f testdevfile ; ls /dev/label > > > > 2. > > I think that is one advantage of glabel: It does not care for the device > > name, but it just tastes the last sector of each device, as soon as it > > shows up (e. g. after the reboot). > > > > > thank's Arne, I will give it a try on a test machine and give you a feedback > on it Also this has nothing to do with FreeBSD performance so followups should be elsewhere. Kris From owner-freebsd-performance@FreeBSD.ORG Fri Jun 1 23:12:54 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 6CA8B16A41F for ; Fri, 1 Jun 2007 23:12:54 +0000 (UTC) (envelope-from amesbury@umn.edu) Received: from mta-a2.tc.umn.edu (mta-a2.tc.umn.edu [134.84.119.206]) by mx1.freebsd.org (Postfix) with ESMTP id 427FF13C46C for ; Fri, 1 Jun 2007 23:12:54 +0000 (UTC) (envelope-from amesbury@umn.edu) Received: from paulaner.oitsec.umn.edu (paulaner.oitsec.umn.edu [160.94.247.212]) by mta-a2.tc.umn.edu (UMN smtpd) with ESMTP for ; Fri, 1 Jun 2007 17:57:47 -0500 (CDT) X-Umn-Remote-Mta: [N] paulaner.oitsec.umn.edu [160.94.247.212] #+LO+TS+AU Message-ID: <4660A46B.3010003@umn.edu> Date: Fri, 01 Jun 2007 17:57:47 -0500 From: Alan Amesbury User-Agent: Thunderbird 2.0.0.0 (X11/20070530) MIME-Version: 1.0 To: freebsd-performance@freebsd.org Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Subject: Intel PRO/10GbE CX4? General 10Gb tips? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 01 Jun 2007 23:12:54 -0000 ixgb(4) on 6.2-RELEASE-p5 says that the driver supports adapters based on the 82597EX controller. The 10GbE CX4 uses the same chip, but isn't in the list of explicitly listed hardware in the manpage. A quick peek at the driver in the source tree shows that it does some hardware-specific checks for the SR adapter, because they "support two different types of XPAK optics....." (ixgb_hw.c). There's no mention of the CX4, and the dates on the source suggest they're from '04. I see Intel has an updated version of this driver from '06 which explicitly lists the CX4. Are any of you familiar with these cards and, if so, do you know if the 6.2-RELEASE or later drivers will work with the CX4? On a related note, do you have any suggestions for 10Gb hardware? I lean towards the Intel stuff because, at least with the 1Gb stuff (and aside from the weird and atypical em(4) problems earlier in 6-RELEASE), the Intel stuff just works. That said, I'm open to suggestions for other 10Gb stuff if you've got them. Besides the tuning stuff discussed here before (around mid-December, 2006), are there knobs that I should tweak for ridiculously fast packet capture? As always, thanks in advance! -- Alan Amesbury OIT Security and Assurance University of Minnesota From owner-freebsd-performance@FreeBSD.ORG Sat Jun 2 00:36:39 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 0D32B16A421 for ; Sat, 2 Jun 2007 00:36:39 +0000 (UTC) (envelope-from kip.macy@gmail.com) Received: from ug-out-1314.google.com (ug-out-1314.google.com [66.249.92.169]) by mx1.freebsd.org (Postfix) with ESMTP id 769C813C455 for ; Sat, 2 Jun 2007 00:36:38 +0000 (UTC) (envelope-from kip.macy@gmail.com) Received: by ug-out-1314.google.com with SMTP id u2so378273uge for ; Fri, 01 Jun 2007 17:36:37 -0700 (PDT) DKIM-Signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=jhHL42IKp9ESgaReE2yfVUTpJPYkvZp+1kjTcyR8+x6lb6s578XU/IPesKISqCEeVN5RJivMElT+MITLhGEeNwvjiblehuzxFHT0tfPo3CcbIMzbhE8fjA9Qx+kGrqk+eA4x3oIO37qFJ6TxRjYmuatsh5wKupEXUHgeZfeLySU= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=OJCqs6gTSJ246OhfM66Mz0THgwkVZZaeuaPq3IIZlqVvX3lpChwjU3WF/vmH8hULNEVHXHUZC9uMVjRFcHdDRlK1XaSbm2re8sHiadNcnDcmHRpEvJjSvz8dSD8ePI3A47wzeYy8sYG7kNSLrwsvCcnpuoqmT7eN1elLqvzQ3Cw= Received: by 10.78.134.12 with SMTP id h12mr1440537hud.1180742898187; Fri, 01 Jun 2007 17:08:18 -0700 (PDT) Received: by 10.78.107.13 with HTTP; Fri, 1 Jun 2007 17:08:18 -0700 (PDT) Message-ID: Date: Fri, 1 Jun 2007 17:08:18 -0700 From: "Kip Macy" To: "Alan Amesbury" In-Reply-To: <4660A46B.3010003@umn.edu> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <4660A46B.3010003@umn.edu> Cc: freebsd-performance@freebsd.org Subject: Re: Intel PRO/10GbE CX4? General 10Gb tips? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 02 Jun 2007 00:36:39 -0000 My understanding is that ixgb in CVS is quite far behind the vendor driver. Unfortunately, for high performance 10GigE on FreeBSD the only options worth looking at the moment are cxgb (Chelsio T3) and mxge (Myricom). Chelsio is selling eval kits right now: http://www.chelsio.com/evalkits.php Myricom nominally sells their cards directly, but they've dropped both a support request and a PO request from me, so I don't think its worthwhile contacting them directly unless you're a large customer. Bell Micro sells both the Chelsio and the Myricom cards. -Kip On 6/1/07, Alan Amesbury wrote: > ixgb(4) on 6.2-RELEASE-p5 says that the driver supports adapters based > on the 82597EX controller. The 10GbE CX4 uses the same chip, but isn't > in the list of explicitly listed hardware in the manpage. A quick peek > at the driver in the source tree shows that it does some > hardware-specific checks for the SR adapter, because they "support two > different types of XPAK optics....." (ixgb_hw.c). There's no mention of > the CX4, and the dates on the source suggest they're from '04. I see > Intel has an updated version of this driver from '06 which explicitly > lists the CX4. > > Are any of you familiar with these cards and, if so, do you know if the > 6.2-RELEASE or later drivers will work with the CX4? > > On a related note, do you have any suggestions for 10Gb hardware? I > lean towards the Intel stuff because, at least with the 1Gb stuff (and > aside from the weird and atypical em(4) problems earlier in 6-RELEASE), > the Intel stuff just works. That said, I'm open to suggestions for > other 10Gb stuff if you've got them. > > Besides the tuning stuff discussed here before (around mid-December, > 2006), are there knobs that I should tweak for ridiculously fast packet > capture? > > As always, thanks in advance! > > > -- > Alan Amesbury > OIT Security and Assurance > University of Minnesota > _______________________________________________ > 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 Sat Jun 2 01:20:49 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id D26D916A41F for ; Sat, 2 Jun 2007 01:20:49 +0000 (UTC) (envelope-from killing@multiplay.co.uk) Received: from multiplay.co.uk (core6.multiplay.co.uk [85.236.96.23]) by mx1.freebsd.org (Postfix) with ESMTP id 6FCD413C48C for ; Sat, 2 Jun 2007 01:20:48 +0000 (UTC) (envelope-from killing@multiplay.co.uk) X-Spam-Checker-Version: SpamAssassin 3.1.5 (2006-08-29) on core6.multiplay.co.uk X-Spam-Level: X-Spam-Status: No, score=-14.7 required=6.0 tests=BAYES_00, USER_IN_WHITELIST, USER_IN_WHITELIST_TO autolearn=ham version=3.1.5 Received: from r2d2 ([212.135.219.182]) by multiplay.co.uk (multiplay.co.uk [85.236.96.23]) (MDaemon PRO v9.5.4) with ESMTP id md50003877518.msg for ; Sat, 02 Jun 2007 02:16:30 +0100 Message-ID: <01e801c7a4b3$a4bcb400$b6db87d4@multiplay.co.uk> From: "Steven Hartland" To: "Kip Macy" , "Alan Amesbury" References: <4660A46B.3010003@umn.edu> Date: Sat, 2 Jun 2007 02:16:26 +0100 MIME-Version: 1.0 Content-Type: text/plain; format=flowed; charset="iso-8859-1"; reply-type=response Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2900.3028 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3028 X-MDRemoteIP: 212.135.219.182 X-Return-Path: killing@multiplay.co.uk X-Envelope-From: killing@multiplay.co.uk X-MDaemon-Deliver-To: freebsd-performance@freebsd.org X-Spam-Processed: multiplay.co.uk, Sat, 02 Jun 2007 02:16:31 +0100 X-MDAV-Processed: multiplay.co.uk, Sat, 02 Jun 2007 02:16:31 +0100 Cc: freebsd-performance@freebsd.org Subject: Re: Intel PRO/10GbE CX4? General 10Gb tips? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 02 Jun 2007 01:20:49 -0000 Wasn't Jack Vogel (Intel?) only talking the other day about committing a new 10Gb Intel driver. The "New driver coming soon" thread on current / net. Steve ================================================ This e.mail is private and confidential between Multiplay (UK) Ltd. and the person or entity to whom it is addressed. In the event of misdirection, the recipient is prohibited from using, copying, printing or otherwise disseminating it or any information contained in it. In the event of misdirection, illegible or incomplete transmission please telephone +44 845 868 1337 or return the E.mail to postmaster@multiplay.co.uk. From owner-freebsd-performance@FreeBSD.ORG Sat Jun 2 01:41:05 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 311A116A421 for ; Sat, 2 Jun 2007 01:41:05 +0000 (UTC) (envelope-from kip.macy@gmail.com) Received: from ug-out-1314.google.com (ug-out-1314.google.com [66.249.92.174]) by mx1.freebsd.org (Postfix) with ESMTP id BDEC913C45A for ; Sat, 2 Jun 2007 01:41:04 +0000 (UTC) (envelope-from kip.macy@gmail.com) Received: by ug-out-1314.google.com with SMTP id u2so384894uge for ; Fri, 01 Jun 2007 18:41:03 -0700 (PDT) DKIM-Signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=gzcXtXP5AdSHLf1nFurZKnRPAfexBptw2Ukl9jK5bHDE2R+9YPx5SxxA1NcaKaS4dpl4LpRR5WI17fgx8BTWQsNzFz/5BNRRPPLdv6FMVUZ/JAk7FZ5uuu7e35wlu2kI9LzXm4NO+dYdK+M79ojb4kKTqNsPIgnSHWzIZzaTFa8= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=Zf5yU+YN6lFzh/u+wu4pVoti/pB4Ln8hPsNVu9fmYmqNE3m+OlRSfV0nlA1XV5c84RzQleQBdE/EjzzVVBESNDexoh8Rvz93plHVv+Lr2ttbVhy4kjBbhW1nsMkVdSNBG2zgymBIfvHsvVJORXysut7W8PksCn4tGHUshR6+AWc= Received: by 10.78.81.20 with SMTP id e20mr1417730hub.1180748463665; Fri, 01 Jun 2007 18:41:03 -0700 (PDT) Received: by 10.78.107.13 with HTTP; Fri, 1 Jun 2007 18:41:03 -0700 (PDT) Message-ID: Date: Fri, 1 Jun 2007 18:41:03 -0700 From: "Kip Macy" To: "Steven Hartland" In-Reply-To: <01e801c7a4b3$a4bcb400$b6db87d4@multiplay.co.uk> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <4660A46B.3010003@umn.edu> <01e801c7a4b3$a4bcb400$b6db87d4@multiplay.co.uk> Cc: freebsd-performance@freebsd.org, Alan Amesbury Subject: Re: Intel PRO/10GbE CX4? General 10Gb tips? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 02 Jun 2007 01:41:05 -0000 On 6/1/07, Steven Hartland wrote: > Wasn't Jack Vogel (Intel?) only talking the other day about > committing a new 10Gb Intel driver. > > The "New driver coming soon" thread on current / net. I'm talking about what is in the tree in the moment. I'll have to withhold judgement on ix until we can actually use it. With luck he'll have time to adapt it to FreeBSD. There are a number of other 10GigE drivers cards that are "supported" by FreeBSD, but they take the linux shim approach and tend to be 3x slower. -Kip From owner-freebsd-performance@FreeBSD.ORG Sat Jun 2 12:13:18 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id B32CA16A421 for ; Sat, 2 Jun 2007 12:13:18 +0000 (UTC) (envelope-from hunreal@gmail.com) Received: from an-out-0708.google.com (an-out-0708.google.com [209.85.132.251]) by mx1.freebsd.org (Postfix) with ESMTP id 6FBE613C455 for ; Sat, 2 Jun 2007 12:13:18 +0000 (UTC) (envelope-from hunreal@gmail.com) Received: by an-out-0708.google.com with SMTP id c14so283444anc for ; Sat, 02 Jun 2007 05:13:17 -0700 (PDT) DKIM-Signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=iHgZK3UWhfRqYclPz15y0NGgyXjn03efkb9bVW9l1JGoaXKEeJlONWRianpliOXCcTiThQHBzY41z6MhyGQYH5GV9lNLx1xdcOaVhn/PqKfcG4hjdqqPj2Vm+/QTDRN1TJmJFPYINSF5rtw0xHkL6XP+7c7JV/UPkgFyc6ulGmQ= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=kePH1MDqxSmySQ1xhYeicVZH+ugs23I1ptXPz2+myr8iIhW+CPh9VYW1XO6+Zjww52k6dQN3p7vuRdjQ63B4WjLw3eGnlbk7OqyAFnDyelZTgGXQZqmW5TaRRVacxpftAKslpdJxN83VIqCqfs/TN0SH12EhJW8irM7unnVMdsM= Received: by 10.100.166.14 with SMTP id o14mr1582736ane.1180786397881; Sat, 02 Jun 2007 05:13:17 -0700 (PDT) Received: by 10.100.95.14 with HTTP; Sat, 2 Jun 2007 05:13:17 -0700 (PDT) Message-ID: <9b6b59500706020513x49cf8aecucf8393a98e478221@mail.gmail.com> Date: Sat, 2 Jun 2007 20:13:17 +0800 From: hshh To: "Claus Guttesen" In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <9b6b59500706010310h145a1804ie97e57aa0c5e90a4@mail.gmail.com> Cc: freebsd-performance@freebsd.org Subject: Re: Bad performance while transfer large block size through NFS. X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 02 Jun 2007 12:13:18 -0000 I tried, but not help. Still only 3798612 bytes/sec while block size is 2MBytes. On 6/1/07, Claus Guttesen wrote: > > I set up one NFS server, and mounted on other server by TCP. Servers > > connected with Giga network, and running 6.2-RELEASE. > > > > But I found the performance is very bad while transfering large block size data. > > For example, I use dd on NFS client to test the speed. > > And ``systat -vm 1'' is displayed HDD is most busy, > > Disks da0 > > KB/t 16.00 > > tps 239 > > MB/s 3.73 > > % busy 91 > > > > Can I do some tuning to improve large block size transfer over NFS? > > Try adding read- and write-size 32768, partial /etc/fstab: > > rw,intr,nfsv3,-a=4,-w=32768,-r=32768 > > or adding -r 32768 -w 32768 while mounting from command line. > > -- > regards > Claus > > When lenity and cruelty play for a kingdom, > the gentlest gamester is the soonest winner. > > Shakespeare > From owner-freebsd-performance@FreeBSD.ORG Sat Jun 2 12:31:30 2007 Return-Path: X-Original-To: freebsd-performance@freebsd.org Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 2A64016A469 for ; Sat, 2 Jun 2007 12:31:30 +0000 (UTC) (envelope-from arne_woerner@yahoo.com) Received: from web30313.mail.mud.yahoo.com (web30313.mail.mud.yahoo.com [209.191.69.75]) by mx1.freebsd.org (Postfix) with SMTP id D064F13C45B for ; Sat, 2 Jun 2007 12:31:29 +0000 (UTC) (envelope-from arne_woerner@yahoo.com) Received: (qmail 8724 invoked by uid 60001); 2 Jun 2007 12:31:29 -0000 DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=X-YMail-OSG:Received:Date:From:Subject:To:Cc:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding:Message-ID; b=SCzJYeAbTCIc0b2L//fNByNj1Q9BR/43+DQDJgiJ9oy00YEVZeZvvqhErcjbyqaKCCHLZ5CachvN/DtwZWsGxrwsWQvkN+xYBRPQg9qt44HtktOdx2RA7efr6+QEjpc3gOB9J2OoX4YD3oEvzCVkrIT/xMKOjGaVG0qwMYh1cO4=; X-YMail-OSG: 3BfdspIVM1lBUQahDzTi9FIHwO4lFxjhcQN.QqDT5JjK.91A4kDYAN4I3DmqGGCnff.8mGxBF39pHRVUbji6prQOoWXrkSkinOmEOxWajwj.7zKOauQIcxHyh6Uksw-- Received: from [77.129.177.243] by web30313.mail.mud.yahoo.com via HTTP; Sat, 02 Jun 2007 05:31:29 PDT Date: Sat, 2 Jun 2007 05:31:29 -0700 (PDT) From: Arne "Wörner" To: hshh In-Reply-To: <9b6b59500706020513x49cf8aecucf8393a98e478221@mail.gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Message-ID: <365881.7988.qm@web30313.mail.mud.yahoo.com> Cc: freebsd-performance@freebsd.org Subject: Re: Bad performance while transfer large block size through NFS. X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 02 Jun 2007 12:31:30 -0000 > I set up one NFS server, and mounted on other server by TCP. Servers > connected with Giga network, and running 6.2-RELEASE. > > But I found the performance is very bad while transfering large block > size data. > I dont know if it makes any sense, but at least the following experiment might be fun: 1. With various block sizes (e. g. 32k, 128k, 1m, 2m) on the client box: dd if=/dev/zero of=/dev/null bs=XXX count=1k 2. With various block sizes (e. g. 32k, 128k, 1m, 2m) on the client box: dd if=/dev/zero bs=XXX count=1k | dd of=/path/to/nfs/mount bs=32k 3. With various block sizes (e. g. 32k, 128k, 1m, 2m) on the client box: dd if=/dev/zero bs=XXX count=1k | dd of=/dev/null bs=32k The idea would be, that possibly the client box has problems, when it has to handle big data blocks inside the kernel (in that case part 1 would show the same performance degradation as ur initial test). -Arne ____________________________________________________________________________________ The fish are biting. Get more visitors on your site using Yahoo! Search Marketing. http://searchmarketing.yahoo.com/arp/sponsoredsearch_v2.php