From owner-freebsd-performance@FreeBSD.ORG Mon May 5 04:00:44 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 CBE8E37B401 for ; Mon, 5 May 2003 04:00:44 -0700 (PDT) Received: from mx02.egartech.com (aloha.egartech.com [62.118.81.133]) by mx1.FreeBSD.org (Postfix) with SMTP id 78F9143F3F for ; Mon, 5 May 2003 04:00:42 -0700 (PDT) (envelope-from temik@egartech.com) Received: (qmail 41721 invoked by uid 85); 5 May 2003 11:00:35 -0000 Received: from temik@egartech.com by mx02.egartech.com with qmail-scanner-1.03 (. Clean. Processed in 1.314162 secs); 05 May 2003 11:00:35 -0000 Received: from unknown (HELO turtle.egar.egartech.com) (192.168.8.4) by 0 with SMTP; 5 May 2003 11:00:33 -0000 content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="koi8-r" Content-Transfer-Encoding: quoted-printable X-MimeOLE: Produced By Microsoft Exchange V6.0.6249.0 Date: Mon, 5 May 2003 15:00:35 +0400 Message-ID: <5235EF9BAE6B7F4CB3735789EEF73B29B06A68@turtle.egar.egartech.com> X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: freebsd-performance Digest, Vol 3, Issue Thread-Index: AcMQRuF2q6+FgRV7Q1G+6stn2iYhkwCroGug From: "Artem Tepponen" To: "Terry Lambert" cc: freebsd-performance@freebsd.org Subject: RE: freebsd-performance Digest, Vol 3, Issue 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, 05 May 2003 11:00:45 -0000 > Old People. >=20 > It's the same reason that people think 8% of a 120GB hard drive > is "a lot of space" and refuse to set their free reserve on their > FS's high enough to avoid fragmentation. Should they be asking 'Why FS design is so flawed that it has this requirement' instead? Is there any working FS for FreeBSD that does not have this requirement? Artem From owner-freebsd-performance@FreeBSD.ORG Mon May 5 04:05:02 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 A7B5937B401 for ; Mon, 5 May 2003 04:05:02 -0700 (PDT) Received: from mx02.egartech.com (aloha.egartech.com [62.118.81.133]) by mx1.FreeBSD.org (Postfix) with SMTP id CE82543F3F for ; Mon, 5 May 2003 04:05:00 -0700 (PDT) (envelope-from temik@egartech.com) Received: (qmail 42028 invoked by uid 85); 5 May 2003 11:04:55 -0000 Received: from temik@egartech.com by mx02.egartech.com with qmail-scanner-1.03 (. Clean. Processed in 0.349709 secs); 05 May 2003 11:04:55 -0000 Received: from unknown (HELO turtle.egar.egartech.com) (192.168.8.4) by 0 with SMTP; 5 May 2003 11:04:54 -0000 content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="koi8-r" Content-Transfer-Encoding: quoted-printable X-MimeOLE: Produced By Microsoft Exchange V6.0.6249.0 Date: Mon, 5 May 2003 15:04:57 +0400 Message-ID: <5235EF9BAE6B7F4CB3735789EEF73B29B06A69@turtle.egar.egartech.com> X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: freebsd-performance Digest, Vol 3, Issue 1 Thread-Index: AcMQRmMUFjTxtDgzRZGpJ7FHQ0DSRgCr08uw From: "Artem Tepponen" To: "Terry Lambert" cc: freebsd-performance@freebsd.org Subject: RE: freebsd-performance Digest, Vol 3, Issue 1 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, 05 May 2003 11:05:03 -0000 > Too bad it's not supported, and too bad that, if it was, the > overhead would be too high because there's not VOP to get the > FS block offsets, so you would have to go trouh the FS code to > swap, and it would be much, much slower. Btw, do you have any fresh numbers on hand that can support this = statement? Naive approach whould be comparing CPU time taken and disk latencies that differ by an order of magnitude and conclude that few microseconds eaten by CPU would go unnoticed compared with milliseconds taken by = disk. Artem From owner-freebsd-performance@FreeBSD.ORG Mon May 5 06:29:13 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 D307837B401 for ; Mon, 5 May 2003 06:29:13 -0700 (PDT) Received: from bluhayz.homeunix.org (ip68-106-103-50.nv.nv.cox.net [68.106.103.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id E660443FAF for ; Mon, 5 May 2003 06:29:12 -0700 (PDT) (envelope-from dero@bluhayz.homeunix.org) Received: from bluhayz.homeunix.org (localhost [127.0.0.1]) h45DUUbG001595 for ; Mon, 5 May 2003 09:30:31 -0400 (EDT) (envelope-from dero@bluhayz.homeunix.org) Received: (from nobody@localhost) by bluhayz.homeunix.org (8.12.8p1/8.12.8/Submit) id h45DUUlq001594; Mon, 5 May 2003 09:30:30 -0400 (EDT) From: agent dero X-Authentication-Warning: bluhayz.homeunix.org: nobody set sender to dero@bluhayz.homeunix.org using -f Received: from 172.182.185.39 (SquirrelMail authenticated user dero) by bluhayz.homeunix.org with HTTP; Mon, 5 May 2003 09:30:30 -0400 (EDT) Message-ID: <3801.172.182.185.39.1052141430.squirrel@bluhayz.homeunix.org> Date: Mon, 5 May 2003 09:30:30 -0400 (EDT) To: freebsd-performance@freebsd.org User-Agent: SquirrelMail/1.4.0 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 X-Priority: 3 Importance: Normal Subject: Real and availible RAM 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, 05 May 2003 13:29:14 -0000 I have been looking through the kernel boot messages in /var/log while working on some custom kernel compile work, and I came across something that I think is very interesting, but doesn't make sense. real memory = 100663296 (98304K bytes) avail memory = 94580736 (92364K bytes) This tells me that FreeBSD recognizes my 98MB of RAM, but it only uses 92MB? Are the 6MB of RAM that are left getting shafted, and just using power, but not being addressed by FreeBSD? Does this slow down my machine at all, I mean, is there a percentage to this? Where only x% of 100% RAM is availible or usable? Also, I use phpSysInfo to judge the status of most of my remote servers, and it shows the caches and buffers as part of the whole chunk of used RAM, so at one point it can be up to 95% of the RAM. But then 10 minutes later it will have dropped back down to 50% or so, showing that the buffers were somehow cleared? Is there anyway to do this manually? From owner-freebsd-performance@FreeBSD.ORG Mon May 5 07:31:08 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 A1A0C37B401 for ; Mon, 5 May 2003 07:31:08 -0700 (PDT) Received: from heron.mail.pas.earthlink.net (heron.mail.pas.earthlink.net [207.217.120.189]) by mx1.FreeBSD.org (Postfix) with ESMTP id 2A4FB43F3F for ; Mon, 5 May 2003 07:31:08 -0700 (PDT) (envelope-from tlambert2@mindspring.com) Received: from pool0250.cvx22-bradley.dialup.earthlink.net ([209.179.198.250] helo=mindspring.com) by heron.mail.pas.earthlink.net with asmtp (SSLv3:RC4-MD5:128) (Exim 3.33 #1) id 19Cgzg-00073i-00; Mon, 05 May 2003 07:31:05 -0700 Message-ID: <3EB6755A.BC035326@mindspring.com> Date: Mon, 05 May 2003 07:29:46 -0700 From: Terry Lambert X-Mailer: Mozilla 4.79 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: Artem Tepponen References: <5235EF9BAE6B7F4CB3735789EEF73B29B06A68@turtle.egar.egartech.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-ELNK-Trace: b1a02af9316fbb217a47c185c03b154d40683398e744b8a4a7e498d60fb0a6ffb592ea16c4c3d887666fa475841a1c7a350badd9bab72f9c350badd9bab72f9c cc: freebsd-performance@freebsd.org Subject: Re: freebsd-performance Digest, Vol 3, Issue 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, 05 May 2003 14:31:08 -0000 Artem Tepponen wrote: > > Old People. > > > > It's the same reason that people think 8% of a 120GB hard drive > > is "a lot of space" and refuse to set their free reserve on their > > FS's high enough to avoid fragmentation. > > Should they be asking 'Why FS design is so flawed > that it has this requirement' instead? Not really. They should be asking "Why do I think this is an abnormal amount of overhead compared to, say, the second of the two journal files NT maintains?". > Is there any working FS for FreeBSD that does not have this requirement? There are several, including a number of Linux-originated ones; however, lacking the requirement, they are all subject to internal fragmentation and needing a defragger or cleaner process in order to deal with the problem. UFS doesn't require this. I wish people would read and understand Donald Knuth's books on algorithms... -- Terry From owner-freebsd-performance@FreeBSD.ORG Mon May 5 07:41:58 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 92EB437B401 for ; Mon, 5 May 2003 07:41:58 -0700 (PDT) Received: from otter3.centtech.com (moat3.centtech.com [207.200.51.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id B2A2B43F75 for ; Mon, 5 May 2003 07:41:57 -0700 (PDT) (envelope-from anderson@centtech.com) Received: from centtech.com (electron.centtech.com [204.177.173.173]) by otter3.centtech.com (8.12.3/8.12.3) with ESMTP id h45Efu56083090 for ; Mon, 5 May 2003 09:41:56 -0500 (CDT) (envelope-from anderson@centtech.com) Message-ID: <3EB67822.3070802@centtech.com> Date: Mon, 05 May 2003 09:41:38 -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: freebsd-performance@freebsd.org Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Subject: NAT performance tweaks 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, 05 May 2003 14:41:58 -0000 Does anyone have any tweaks they apply to NAT firewalls that pass a lot of connections through them? Here's the ony tweak I have in place already, but I'm not sure they're needed yet (or if there are any tweaks needed at all): sysctl kern.ipc.somaxconn=8192 Most connections being passed through this box are http requests, mail, ftp, and ssh connections. Any hints? Eric -- ------------------------------------------------------------------ Eric Anderson Systems Administrator Centaur Technology Attitudes are contagious, is yours worth catching? ------------------------------------------------------------------ From owner-freebsd-performance@FreeBSD.ORG Mon May 5 08:44:28 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 9D72337B401 for ; Mon, 5 May 2003 08:44:28 -0700 (PDT) Received: from hotmail.com (bay1-f72.bay1.hotmail.com [65.54.245.72]) by mx1.FreeBSD.org (Postfix) with ESMTP id 2BE0143FA3 for ; Mon, 5 May 2003 08:44:28 -0700 (PDT) (envelope-from rob_macgregor@hotmail.com) Received: from mail pickup service by hotmail.com with Microsoft SMTPSVC; Mon, 5 May 2003 08:44:28 -0700 Received: from 80.192.46.45 by by1fd.bay1.hotmail.msn.com with HTTP; Mon, 05 May 2003 15:44:28 GMT X-Originating-IP: [80.192.46.45] X-Originating-Email: [rob_macgregor@hotmail.com] From: "Rob MacGregor" To: freebsd-performance@freebsd.org Date: Mon, 05 May 2003 15:44:28 +0000 Mime-Version: 1.0 Content-Type: text/plain; format=flowed Message-ID: X-OriginalArrivalTime: 05 May 2003 15:44:28.0131 (UTC) FILETIME=[363AD330:01C3131D] Subject: Re: Real and availible RAM X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: KEEP_TRAFFIC_ON_THE_LIST@dev.null List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 05 May 2003 15:44:28 -0000 >From: agent dero > >I have been looking through the kernel boot messages in /var/log while >working on some custom kernel compile work, and I came across something >that I think is very interesting, but doesn't make sense. >real memory = 100663296 (98304K bytes) >avail memory = 94580736 (92364K bytes) I could be wrong, but if FreeBSD is anything like Solaris what it's telling you is that there is 96 MB of RAM installed and that once FreeBSD's kernel is loaded there is 90 MB of RAM left. So, the FreeBSD kernel is using some 6 MB. >Also, I use phpSysInfo to judge the status of most of my remote servers, >and it shows the caches and buffers as part of the whole chunk of used >RAM, so at one point it can be up to 95% of the RAM. But then 10 minutes >later it will have dropped back down to 50% or so, showing that the >buffers were somehow cleared? Is there anyway to do this manually? I'm pretty sure the full details are lurking in the handbook, however from what I understand the caches and buffers are using RAM that's otherwise idle. If something needs that RAM then the caches/buffers shrink. If there's lots of free RAM then they grow. This is a Good Thing not a bad thing. Please DO NOT send me ANY email directly unless it's a privacy issue. Reply-to mangled to assist those who don't read the above. -- Rob | What part of "no" was it you didn't understand? _________________________________________________________________ Worried what your kids see online? Protect them better with MSN 8 http://join.msn.com/?page=features/parental&pgmarket=en-gb&XAPID=186&DI=1059 From owner-freebsd-performance@FreeBSD.ORG Mon May 5 09:28:01 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 6CF0337B401 for ; Mon, 5 May 2003 09:28:01 -0700 (PDT) Received: from cultdeadsheep.org (charon.cultdeadsheep.org [80.65.226.72]) by mx1.FreeBSD.org (Postfix) with ESMTP id D6E8843FB1 for ; Mon, 5 May 2003 09:27:59 -0700 (PDT) (envelope-from sheep.killer@cultdeadsheep.org) Received: (qmail 74760 invoked from network); 5 May 2003 16:27:58 -0000 Received: from unknown (HELO lucifer.cultdeadsheep.org) (192.168.0.2) by goofy.cultdeadsheep.org with SMTP; 5 May 2003 16:27:58 -0000 Date: Mon, 5 May 2003 18:27:56 +0200 From: Clement Laforet To: freebsd-performance@freebsd.org Message-Id: <20030505182756.093fb1c3.sheep.killer@cultdeadsheep.org> In-Reply-To: <3EB67822.3070802@centtech.com> References: <3EB67822.3070802@centtech.com> Organization: tH3 cUlt 0f tH3 d3@d sH33p X-Mailer: Sylpheed version 0.8.11 (GTK+ 1.2.10; i386-portbld-freebsd4.8) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Subject: Re: NAT performance tweaks 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, 05 May 2003 16:28:01 -0000 On Mon, 05 May 2003 09:41:38 -0500 Eric Anderson wrote: > Does anyone have any tweaks they apply to NAT firewalls that pass a > lot of connections through them? Here's the ony tweak I have in place > already, but I'm not sure they're needed yet (or if there are any > tweaks needed at all): which NAT solution do you use ? > sysctl kern.ipc.somaxconn=8192 NAT'ing (except for natd which uses IPDIVERT (but not more than 3)) doesn't use socket to translate packets. Generally, packets are tagged by firewall control software and translated within the IP stack (at leat in kernel land). clem From owner-freebsd-performance@FreeBSD.ORG Mon May 5 10:35:16 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 CE96237B401 for ; Mon, 5 May 2003 10:35:16 -0700 (PDT) Received: from otter3.centtech.com (moat3.centtech.com [207.200.51.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 0543B43F3F for ; Mon, 5 May 2003 10:35:16 -0700 (PDT) (envelope-from anderson@centtech.com) Received: from centtech.com (electron.centtech.com [204.177.173.173]) by otter3.centtech.com (8.12.3/8.12.3) with ESMTP id h45HZC56006867; Mon, 5 May 2003 12:35:12 -0500 (CDT) (envelope-from anderson@centtech.com) Message-ID: <3EB6A0BF.1040803@centtech.com> Date: Mon, 05 May 2003 12:34:55 -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: Clement Laforet References: <3EB67822.3070802@centtech.com> <20030505182756.093fb1c3.sheep.killer@cultdeadsheep.org> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org Subject: Re: NAT performance tweaks 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, 05 May 2003 17:35:17 -0000 Clement Laforet wrote: > On Mon, 05 May 2003 09:41:38 -0500 > Eric Anderson wrote: > > >>Does anyone have any tweaks they apply to NAT firewalls that pass a >>lot of connections through them? Here's the ony tweak I have in place >>already, but I'm not sure they're needed yet (or if there are any >>tweaks needed at all): > > > which NAT solution do you use ? IPNAT and ipfilter.. >>sysctl kern.ipc.somaxconn=8192 > > > NAT'ing (except for natd which uses IPDIVERT (but not more than 3)) > doesn't use socket to translate packets. > Generally, packets are tagged by firewall control software and > translated within the IP stack (at leat in kernel land). Oh yea, that's right.. So can you think of any kernel or other tweaks to be done, to ensure optimal usage of the machine in this environment? What about mail coming in/out of the machine? I do a fair amount of mail through it (out through NAT, in through Sendmail) also.. Eric -- ------------------------------------------------------------------ Eric Anderson Systems Administrator Centaur Technology Attitudes are contagious, is yours worth catching? ------------------------------------------------------------------ From owner-freebsd-performance@FreeBSD.ORG Mon May 5 11:08:04 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 38DE837B401 for ; Mon, 5 May 2003 11:08:04 -0700 (PDT) Received: from xy.hartford.edu (xy.hartford.edu [137.49.19.55]) by mx1.FreeBSD.org (Postfix) with ESMTP id 2F00843F93 for ; Mon, 5 May 2003 11:08:02 -0700 (PDT) (envelope-from knr@xy.hartford.edu) Received: from xy.hartford.edu (knr@localhost.hartford.edu [127.0.0.1]) by xy.hartford.edu (8.12.9/8.12.6) with ESMTP id h45I81Jx000394 for ; Mon, 5 May 2003 14:08:01 -0400 (EDT) (envelope-from knr@xy.hartford.edu) From: "Kyle Rollin" To: freebsd-performance@freebsd.org Date: Mon, 5 May 2003 13:08:01 -0500 Message-Id: <20030505180224.M13866@xy.hartford.edu> In-Reply-To: <3801.172.182.185.39.1052141430.squirrel@bluhayz.homeunix.org> References: <3801.172.182.185.39.1052141430.squirrel@bluhayz.homeunix.org> X-Mailer: Open WebMail 2.00 20030416 X-OriginatingIP: 137.49.19.54 (knr) MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Subject: Re: Real and availible RAM 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, 05 May 2003 18:08:04 -0000 On Mon, 5 May 2003 09:30:30 -0400 (EDT), agent dero wrote > I have been looking through the kernel boot messages in /var/log > while working on some custom kernel compile work, and I came across something > that I think is very interesting, but doesn't make sense. > real memory = 100663296 (98304K bytes) > avail memory = 94580736 (92364K bytes) > > This tells me that FreeBSD recognizes my 98MB of RAM, but it only > uses 92MB? Are the 6MB of RAM that are left getting shafted, and > just using power, but not being addressed by FreeBSD? Does this slow > down my machine at all, I mean, is there a percentage to this? Where > only x% of 100% RAM is availible or usable? > If you look at the way x86 architecture is designed (and somebody else can feel free to correct me if I'm wrong), but system memory is often used in the caching/shadowing of BIOS. This is where a lot of system memory often goes before the OS is loaded - also, as Rob said, the kernel itself will take up memory before the rest of the OS is booted. If you're concerned that you might run out of memory, RAM is cheap - adding a stick of 128MB will greatly reduce that risk :) -Kyle Rollin From owner-freebsd-performance@FreeBSD.ORG Mon May 5 11:12:36 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 2A12A37B401 for ; Mon, 5 May 2003 11:12:36 -0700 (PDT) Received: from otter3.centtech.com (moat3.centtech.com [207.200.51.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 77A3643F85 for ; Mon, 5 May 2003 11:12:35 -0700 (PDT) (envelope-from anderson@centtech.com) Received: from centtech.com (electron.centtech.com [204.177.173.173]) by otter3.centtech.com (8.12.3/8.12.3) with ESMTP id h45ICY56011316; Mon, 5 May 2003 13:12:35 -0500 (CDT) (envelope-from anderson@centtech.com) Message-ID: <3EB6A981.1030504@centtech.com> Date: Mon, 05 May 2003 13:12:17 -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: Kyle Rollin References: <3801.172.182.185.39.1052141430.squirrel@bluhayz.homeunix.org> <20030505180224.M13866@xy.hartford.edu> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org Subject: Re: Real and availible RAM 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, 05 May 2003 18:12:36 -0000 Kyle Rollin wrote: > On Mon, 5 May 2003 09:30:30 -0400 (EDT), agent dero wrote > >>I have been looking through the kernel boot messages in /var/log >>while working on some custom kernel compile work, and I came across > > something > >>that I think is very interesting, but doesn't make sense. >>real memory = 100663296 (98304K bytes) >>avail memory = 94580736 (92364K bytes) >> >>This tells me that FreeBSD recognizes my 98MB of RAM, but it only >>uses 92MB? Are the 6MB of RAM that are left getting shafted, and >>just using power, but not being addressed by FreeBSD? Does this slow >>down my machine at all, I mean, is there a percentage to this? Where >>only x% of 100% RAM is availible or usable? >> > > > > > If you look at the way x86 architecture is designed (and somebody else can > feel free to correct me if I'm wrong), but system memory is often used in > the caching/shadowing of BIOS. This is where a lot of system memory often > goes before the OS is loaded - also, as Rob said, the kernel itself will > take up memory before the rest of the OS is booted. I believe that the BIOS steals the ram prior to the OS booting, so that the box does not show that as part of the total ram.. Like: CPU: VIA/IDT Unknown (1066.16-MHz 686-class CPU) Origin = "CentaurHauls" Id = 0x693 Stepping = 3 Features=0x380b13d real memory = 234815488 (223 MB) avail memory = 221167616 (210 MB) Eric -- ------------------------------------------------------------------ Eric Anderson Systems Administrator Centaur Technology Attitudes are contagious, is yours worth catching? ------------------------------------------------------------------ From owner-freebsd-performance@FreeBSD.ORG Mon May 5 11:37:43 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 3F1C837B401 for ; Mon, 5 May 2003 11:37:43 -0700 (PDT) Received: from perrin.int.nxad.com (internal.ext.nxad.com [69.1.70.251]) by mx1.FreeBSD.org (Postfix) with ESMTP id C895543FDF for ; Mon, 5 May 2003 11:37:42 -0700 (PDT) (envelope-from sean@perrin.int.nxad.com) Received: by perrin.int.nxad.com (Postfix, from userid 1001) id 7F40D21058; Mon, 5 May 2003 11:37:41 -0700 (PDT) Date: Mon, 5 May 2003 11:37:41 -0700 From: Sean Chittenden To: "Thomas Krause (Webmatic)" Message-ID: <20030505183741.GH94932@perrin.int.nxad.com> References: <3EAE2AB9.4030408@webmatic.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3EAE2AB9.4030408@webmatic.de> User-Agent: Mutt/1.4i X-PGP-Key: finger seanc@FreeBSD.org X-PGP-Fingerprint: 3849 3760 1AFE 7B17 11A0 83A6 DD99 E31F BC84 B341 X-Web-Homepage: http://sean.chittenden.org/ cc: freebsd-performance@freebsd.org Subject: Re: apache2 tuning 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, 05 May 2003 18:37:43 -0000 > I want to build a new webserver (dual xeon with 4 GB RAM). The > server provides mostly dynamic php-pages. In the ports Makefile > there are compile option like WITH_THREADS and WITH_MPM (which > includes WITH_THREADS). These are useful options for a production > machine? Also, is it useful to compile the kernel with > ACCEPT_FILTER_DATA and ACCEPT_FILTER_HTTP? And if yes, is apache > compiled with accept filter by default? Having the accept filters turned on in the kernel is a huge win for Apache. See the tuning(7) man page for other details of interest. As for which MPM, last I heard UNIX folk were better off using the pre-fork MPM and not the threaded MPM. This may have changed, but I'd bet dime to dollar you'll get better stability out of the pre-fork but a smaller memory footprint out of the threaded MPM. -sc -- Sean Chittenden From owner-freebsd-performance@FreeBSD.ORG Mon May 5 11:54:55 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 EFF8E37B401 for ; Mon, 5 May 2003 11:54:55 -0700 (PDT) Received: from smtpout.mac.com (A17-250-248-87.apple.com [17.250.248.87]) by mx1.FreeBSD.org (Postfix) with ESMTP id 1638543F75 for ; Mon, 5 May 2003 11:54:55 -0700 (PDT) (envelope-from cswiger@mac.com) Received: from mac.com (smtpin08-en2 [10.13.10.153]) by smtpout.mac.com (Xserve/MantshX 2.0) with ESMTP id h45IsEIi009835 for ; Mon, 5 May 2003 11:54:54 -0700 (PDT) Received: from mac.com (dpvc-68-161-244-25.ny325.east.verizon.net [68.161.244.25]) (authenticated bits=0) by mac.com (Xserve/MantshX 2.0) with ESMTP id h45IjMDY022962 for ; Mon, 5 May 2003 11:45:23 -0700 (PDT) Date: Mon, 5 May 2003 14:47:07 -0400 Content-Type: text/plain; charset=US-ASCII; format=flowed Mime-Version: 1.0 (Apple Message framework v552) From: Charles Swiger To: freebsd-performance@freebsd.org Content-Transfer-Encoding: 7bit In-Reply-To: <5235EF9BAE6B7F4CB3735789EEF73B29B06A68@turtle.egar.egartech.com> Message-Id: X-Mailer: Apple Mail (2.552) Subject: FFS and minfree... 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, 05 May 2003 18:54:56 -0000 On Monday, May 5, 2003, at 07:00 AM, Artem Tepponen wrote: [ ... ] > Should they be asking 'Why FS design is so flawed > that it has this requirement' instead? No. > Is there any working FS for FreeBSD that does not have this > requirement? The FFS doesn't have a requirement that you reserve any free space, although doing so is a reasonable default for most circumstances due to the performance advantages. Do you enjoy having to defragment other filesystem types, by any chance? -- -Chuck From owner-freebsd-performance@FreeBSD.ORG Mon May 5 16:22:05 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 4900537B401 for ; Mon, 5 May 2003 16:22:05 -0700 (PDT) Received: from mail023.syd.optusnet.com.au (mail023.syd.optusnet.com.au [210.49.20.162]) by mx1.FreeBSD.org (Postfix) with ESMTP id DFCAE43FB1 for ; Mon, 5 May 2003 16:22:03 -0700 (PDT) (envelope-from simon@synatech.com.au) Received: from co3021625-a.mckinn1.vic.optushome.com.au (c17681.mckinn1.vic.optusnet.com.au [210.49.221.35])h45NM2l26159 for ; Tue, 6 May 2003 09:22:02 +1000 Received: by co3021625-a.mckinn1.vic.optushome.com.au (Postfix, from userid 1001) id EEA4014A82; Tue, 6 May 2003 09:22:01 +1000 (EST) Date: Tue, 6 May 2003 09:22:01 +1000 From: Simon Lai To: freebsd-performance@freebsd.org Message-ID: <20030506092201.A53586@pobox.com.> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i Sender: sjlai@synatech.com.au X-Mailman-Approved-At: Mon, 05 May 2003 16:27:39 -0700 Subject: 180,000+ kevents - out of memory X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: sjlai@synatech.com.au List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 05 May 2003 23:22:05 -0000 Hi, I want to monitor a large number of sockets using kevents. I get an out of memory error when I try to submit more than 180,000 kevents. What kernel tunables will allow me to increase the number of kevents that I can allocate? regs Simon From owner-freebsd-performance@FreeBSD.ORG Mon May 5 19:57:42 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 5F02137B401 for ; Mon, 5 May 2003 19:57:42 -0700 (PDT) Received: from heron.mail.pas.earthlink.net (heron.mail.pas.earthlink.net [207.217.120.189]) by mx1.FreeBSD.org (Postfix) with ESMTP id C613B43F3F for ; Mon, 5 May 2003 19:57:41 -0700 (PDT) (envelope-from tlambert2@mindspring.com) Received: from pool0018.cvx21-bradley.dialup.earthlink.net ([209.179.192.18] helo=mindspring.com) by heron.mail.pas.earthlink.net with asmtp (SSLv3:RC4-MD5:128) (Exim 3.33 #1) id 19CseA-0006mZ-00; Mon, 05 May 2003 19:57:39 -0700 Message-ID: <3EB72454.519DF3E@mindspring.com> Date: Mon, 05 May 2003 19:56:20 -0700 From: Terry Lambert X-Mailer: Mozilla 4.79 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: Artem Tepponen References: <5235EF9BAE6B7F4CB3735789EEF73B29B06A69@turtle.egar.egartech.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-ELNK-Trace: b1a02af9316fbb217a47c185c03b154d40683398e744b8a49718f3ba2992e79f9fc08d12e75daf2f666fa475841a1c7a350badd9bab72f9c350badd9bab72f9c cc: freebsd-performance@freebsd.org Subject: Re: freebsd-performance Digest, Vol 3, Issue 1 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: Tue, 06 May 2003 02:57:42 -0000 Artem Tepponen wrote: > > Too bad it's not supported, and too bad that, if it was, the > > overhead would be too high because there's not VOP to get the > > FS block offsets, so you would have to go trouh the FS code to > > swap, and it would be much, much slower. > > Btw, do you have any fresh numbers on hand that can support this statement? > Naive approach whould be comparing CPU time taken and disk latencies > that differ by an order of magnitude and conclude that few microseconds > eaten by CPU would go unnoticed compared with milliseconds taken by disk. The FS orders operations; raw disk I/O does not. The FS lays out blocks in files essentially at random; the layout of the blocks in the swap partition is linear. The FS must obey POSIX semantics about access and modification times; raw disk I/O does not. The FS enforces read-before-write on non-page aligned whole page access. We aren't talking about CPU time here, we are talking about operational delay overhead, seek overhead, and a doubling of the addition of a write operation per read or write access to the file, etc.. You can't tell me that twice the I/O... potentially twice the I/O... is a CPU issue. Even with the optimization I suggested, of getting a physical block list, and using that against the raw device (essentially the same pig-trick that the FreeBSD NTFS uses to rewrite NTFS files contents, so long as the size never changes), there's still an additional indirection through a blocklist to convert a physically discontiguous block array into a logically contiguous one, and there's still the fact that it has to seek all over the disk to access those blocks, and it can't use bulk transfer in the driver or predictive read-ahead in the VM system. Add to this that you can't dump to a swap device created this way: at crash time, you cannot risk extending the file, so it would have to be pre-allocated large enough, and you could not trust the block conversion list was not corrupted by whatever caused the panic, and where you write is not limited by a simple set of block offsets for a region of the disk which is guaranteed to not contain boot-critical or recovery-critical data... ...and you have an overwhelming set of performance limitations not related to CPU utilization. -- Terry From owner-freebsd-performance@FreeBSD.ORG Mon May 5 19:59:28 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 0813737B401 for ; Mon, 5 May 2003 19:59:28 -0700 (PDT) Received: from mailer.neoline.com.br (mailer.neoline.com.br [200.199.96.26]) by mx1.FreeBSD.org (Postfix) with SMTP id 73D5843F75 for ; Mon, 5 May 2003 19:59:26 -0700 (PDT) (envelope-from freebsd@staff.neoline.com.br) Received: (qmail 4084 invoked from network); 6 May 2003 02:59:24 -0000 Received: from pf04.netwaybbs.com.br (HELO AJMNOTEBOOK) (200.199.78.4) by mailer.neoline.com.br with SMTP; 6 May 2003 02:59:24 -0000 Message-ID: <001901c3137b$cc8d66f0$044ec7c8@AJMNOTEBOOK> From: "Alan Jorge Markus" To: "freebsd-performance" Date: Tue, 6 May 2003 00:01:32 -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 Subject: FReebsd 5.0 network performance problem 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: Tue, 06 May 2003 02:59:28 -0000 Hi all! I have two servers, one Freebsd 5.0 running squid and another server with identical hardware configuration, running Freebsd 4.8 and squid, with the same number of requisitions. The problem: I have the same hardware in both servers, with almost the same configuration, but running diferent versions of freebsd, but the machine load in my freebsd running the version 5.0 is 1.40 while the load in the 4.8 server is 0.38. I believe that something in my network configuration maybe wrong. Looking at my 5.0 server: Using "ps axu" i have the proc above: root 19 45.4 0.0 0 12 ?? W 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 82FC737B401 for ; Mon, 5 May 2003 20:19:47 -0700 (PDT) Received: from heron.mail.pas.earthlink.net (heron.mail.pas.earthlink.net [207.217.120.189]) by mx1.FreeBSD.org (Postfix) with ESMTP id F267843FBD for ; Mon, 5 May 2003 20:19:46 -0700 (PDT) (envelope-from tlambert2@mindspring.com) Received: from pool0018.cvx21-bradley.dialup.earthlink.net ([209.179.192.18] helo=mindspring.com) by heron.mail.pas.earthlink.net with asmtp (SSLv3:RC4-MD5:128) (Exim 3.33 #1) id 19CszX-0003Nt-00; Mon, 05 May 2003 20:19:44 -0700 Message-ID: <3EB72987.927E4396@mindspring.com> Date: Mon, 05 May 2003 20:18:31 -0700 From: Terry Lambert X-Mailer: Mozilla 4.79 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: agent dero References: <3801.172.182.185.39.1052141430.squirrel@bluhayz.homeunix.org> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-ELNK-Trace: b1a02af9316fbb217a47c185c03b154d40683398e744b8a4e7a3804c4321f93f7743084de0a2e9002601a10902912494350badd9bab72f9c350badd9bab72f9c cc: freebsd-performance@freebsd.org Subject: Re: Real and availible RAM 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: Tue, 06 May 2003 03:19:47 -0000 agent dero wrote: > I have been looking through the kernel boot messages in /var/log while > working on some custom kernel compile work, and I came across something > that I think is very interesting, but doesn't make sense. > real memory = 100663296 (98304K bytes) > avail memory = 94580736 (92364K bytes) > > This tells me that FreeBSD recognizes my 98MB of RAM, but it only uses 92MB? No. It means that FreeBSD used 6MB of RAM to load itself through locore.s mp_machdep.c, and init_main.c before it got to the point where it enabled the virtual memory system to the point it could account it for you. Also note that "real memory" is not the same as "physical memory" (I frequently add a third line called "phys memory =" to my status display, when I start caring about where RAM is going to enough that I worry about 6M here or there). > Are the 6MB of RAM that are left getting shafted, and just using power, > but not being addressed by FreeBSD? No. You need to read and understand the code that reports these two values, and why it reports a discrepancy. > Does this slow down my machine at all, > I mean, is there a percentage to this? Where only x% of 100% RAM is > availible or usable? Yes and no. You can't reuse any memory used for page tables, so that's "pure overhead". You also can't reuse any memory that has not been marked as swappable. If you allocate memory in the kernel, it will be permanently subtracted from the "avail" (it just doesn't give you another message to the effect of any chanes in "avail", because it would be printf'ing to the console all the time, if it did), since kernel memory is allocated as being type-stable. > Also, I use phpSysInfo to judge the status of most of my remote servers, > and it shows the caches and buffers as part of the whole chunk of used > RAM, so at one point it can be up to 95% of the RAM. But then 10 minutes > later it will have dropped back down to 50% or so, showing that the > buffers were somehow cleared? Is there anyway to do this manually? 1) Use "sync", and wait 2) Kill processes -- Terry From owner-freebsd-performance@FreeBSD.ORG Mon May 5 21:09:55 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 250FC37B401 for ; Mon, 5 May 2003 21:09:55 -0700 (PDT) Received: from heron.mail.pas.earthlink.net (heron.mail.pas.earthlink.net [207.217.120.189]) by mx1.FreeBSD.org (Postfix) with ESMTP id 7CB5743FA3 for ; Mon, 5 May 2003 21:09:54 -0700 (PDT) (envelope-from tlambert2@mindspring.com) Received: from pool0018.cvx21-bradley.dialup.earthlink.net ([209.179.192.18] helo=mindspring.com) by heron.mail.pas.earthlink.net with asmtp (SSLv3:RC4-MD5:128) (Exim 3.33 #1) id 19Ctm3-0003k6-00; Mon, 05 May 2003 21:09:52 -0700 Message-ID: <3EB7353F.373470F2@mindspring.com> Date: Mon, 05 May 2003 21:08:31 -0700 From: Terry Lambert X-Mailer: Mozilla 4.79 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: Eric Anderson References: <3801.172.182.185.39.1052141430.squirrel@bluhayz.homeunix.org> <3EB6A981.1030504@centtech.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-ELNK-Trace: b1a02af9316fbb217a47c185c03b154d40683398e744b8a4426e50e378b293f1d616dcb351543252548b785378294e88350badd9bab72f9c350badd9bab72f9c cc: freebsd-performance@freebsd.org cc: Kyle Rollin Subject: Re: Real and availible RAM 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: Tue, 06 May 2003 04:09:55 -0000 Eric Anderson wrote: > > If you look at the way x86 architecture is designed (and somebody else can > > feel free to correct me if I'm wrong), but system memory is often used in > > the caching/shadowing of BIOS. This is where a lot of system memory often > > goes before the OS is loaded - also, as Rob said, the kernel itself will > > take up memory before the rest of the OS is booted. > > I believe that the BIOS steals the ram prior to the OS booting, so that > the box does not show that as part of the total ram.. Like: Since the BIOS is where FreeBSD obtains the "real memory" number, this is pretty unlikely; the BIOS would just lie to FreeBSD about the number, and be done with it. Also, the 6M is the FreeBSD kernel, page tables, and some misc stuff. You'll note that the "real memory" number is smaller than the expected number for the amount of physical RAM: it already contains the lies by the BIOS. Basically, FreeBSD does not retroactively repair the accounting of the VM system to account for memory consumed before it was up and running -- e.g. it uses pages, and it (in fact) loses pages as well (for example the pages used for the 4K page tables for the kernel prior to the switch over to 4M pages, if you switch to 4M pages, are never returned to the VM system, among many others). The "real" vs. "avail" are a courtesy, between what FreeBSD was told by the BIOS, vs. what is known to the FreeBSD VM. I really suggest anyone who thinks any differently read and understand the code in /sys/i386/i386/machdep.c; specificfically, getmemsize(), where "basemem", "extmem", and Maxmem are calculated, and in the code that gets run in cpu_startup() that prints out the "avail" from ptoa(cnt.v_free_count) -- the VM system's idea of the size of memory available to the kernel. -- Terry From owner-freebsd-performance@FreeBSD.ORG Mon May 5 21:20:51 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 F0EAE37B401 for ; Mon, 5 May 2003 21:20:51 -0700 (PDT) Received: from c7.campus.utcluj.ro (c7.campus.utcluj.ro [193.226.6.226]) by mx1.FreeBSD.org (Postfix) with SMTP id C887943F85 for ; Mon, 5 May 2003 21:20:45 -0700 (PDT) (envelope-from veedee@c7.campus.utcluj.ro) Received: (qmail 84680 invoked by uid 1008); 6 May 2003 04:20:44 -0000 Date: 6 May 2003 07:20:44 +0300 Message-ID: <20030506042044.GA84589@c7.campus.utcluj.ro> From: veedee@c7.campus.utcluj.ro To: "Eric Anderson" References: <3EB67822.3070802@centtech.com> <20030505182756.093fb1c3.sheep.killer@cultdeadsheep.org> <3EB6A0BF.1040803@centtech.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3EB6A0BF.1040803@centtech.com> cc: freebsd-performance@freebsd.org cc: Clement Laforet Subject: Re: NAT performance tweaks 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: Tue, 06 May 2003 04:20:52 -0000 On Mon, May 05, 2003 at 12:34:55PM -0500, Eric Anderson wrote: > Clement Laforet wrote: > >On Mon, 05 May 2003 09:41:38 -0500 > >Eric Anderson wrote: > > > >>Does anyone have any tweaks they apply to NAT firewalls that pass a > >>lot of connections through them? Here's the ony tweak I have in place > >>already, but I'm not sure they're needed yet (or if there are any > >>tweaks needed at all): > > > >which NAT solution do you use ? > > IPNAT and ipfilter.. > > >>sysctl kern.ipc.somaxconn=8192 > > > > > >NAT'ing (except for natd which uses IPDIVERT (but not more than 3)) > >doesn't use socket to translate packets. > >Generally, packets are tagged by firewall control software and > >translated within the IP stack (at leat in kernel land). > > Oh yea, that's right.. So can you think of any kernel or other tweaks to > be done, to ensure optimal usage of the machine in this environment? > What about mail coming in/out of the machine? I do a fair amount of mail > through it (out through NAT, in through Sendmail) also.. If you have a large network behind your NAT server, defining LARGE_NAT in src/contrib/ipfilter/ip_nat.h and src/sys/contrib/ipfilter/netinet/ip_nat.h might help. Don't forget to recompile the kernel and ipfilter. Strange enough, I used to have huge pings (up to 80ms in a totally switched gigabit network) after a few hours of utilization before fiddling with LARGE_NAT. > Eric > > > -- > ------------------------------------------------------------------ > Eric Anderson Systems Administrator Centaur Technology > Attitudes are contagious, is yours worth catching? > ------------------------------------------------------------------ -- | Radu Bogdan Rusu | Network Administrator @ campus.utcluj.ro | | cvsup3.ro/www4.ro.freebsd.org maintainer |->5b736c616d215d<-| | Faculty of Automation & Computer Science @ UTCluj , Romania | |-------------------------------------------------------------| From owner-freebsd-performance@FreeBSD.ORG Mon May 5 23:58:41 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 1CE8937B40C for ; Mon, 5 May 2003 23:58:39 -0700 (PDT) Received: from perrin.int.nxad.com (internal.ext.nxad.com [69.1.70.251]) by mx1.FreeBSD.org (Postfix) with ESMTP id 9E65943F85 for ; Mon, 5 May 2003 23:58:38 -0700 (PDT) (envelope-from sean@perrin.int.nxad.com) Received: by perrin.int.nxad.com (Postfix, from userid 1001) id C3C3720F00; Mon, 5 May 2003 23:58:37 -0700 (PDT) Date: Mon, 5 May 2003 23:58:37 -0700 From: Sean Chittenden To: Alan Jorge Markus Message-ID: <20030506065837.GT94932@perrin.int.nxad.com> References: <001901c3137b$cc8d66f0$044ec7c8@AJMNOTEBOOK> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <001901c3137b$cc8d66f0$044ec7c8@AJMNOTEBOOK> User-Agent: Mutt/1.4i X-PGP-Key: finger seanc@FreeBSD.org X-PGP-Fingerprint: 3849 3760 1AFE 7B17 11A0 83A6 DD99 E31F BC84 B341 X-Web-Homepage: http://sean.chittenden.org/ cc: freebsd-performance Subject: Re: FReebsd 5.0 network performance problem 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: Tue, 06 May 2003 06:58:41 -0000 > I have two servers, one Freebsd 5.0 running squid and another server > with identical hardware configuration, running Freebsd 4.8 and > squid, with the same number of requisitions. > > The problem: > I have the same hardware in both servers, with almost the same > configuration, but running diferent versions of freebsd, but the machine > load in my freebsd running the version 5.0 is 1.40 while the load in the 4.8 > server is 0.38. > > I believe that something in my network configuration maybe wrong. > Looking at my 5.0 server: Don't use 5.0 for production. If you are going to put anything in production, use a _very_ recent -CURRENT. Even then, 5.0 isn't as fast as the -STABLE series at the moment because it's still in an awkward transition between locking via giant and fine grain locking. If you'd like to help figure out performance problems for -CURRENT however, have you tried running any kernel profiling to see where the kernel is spending it's cycles? -sc -- Sean Chittenden From owner-freebsd-performance@FreeBSD.ORG Tue May 6 04:39:17 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 3AB8037B404 for ; Tue, 6 May 2003 04:39:17 -0700 (PDT) Received: from angelica.unixdaemons.com (angelica.unixdaemons.com [209.148.64.135]) by mx1.FreeBSD.org (Postfix) with ESMTP id 1666043F75 for ; Tue, 6 May 2003 04:39:16 -0700 (PDT) (envelope-from hiten@angelica.unixdaemons.com) Received: from angelica.unixdaemons.com (hiten@localhost.unixdaemons.com [127.0.0.1])h46BdDww032261; Tue, 6 May 2003 07:39:13 -0400 (EDT) Received: (from hiten@localhost) by angelica.unixdaemons.com (8.12.9/8.12.1/Submit) id h46BdCAt032260; Tue, 6 May 2003 07:39:12 -0400 (EDT) (envelope-from hiten) Date: Tue, 6 May 2003 07:39:12 -0400 From: Hiten Pandya To: Sean Chittenden Message-ID: <20030506113912.GA31085@unixdaemons.com> References: <3EAE2AB9.4030408@webmatic.de> <20030505183741.GH94932@perrin.int.nxad.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20030505183741.GH94932@perrin.int.nxad.com> User-Agent: Mutt/1.4.1i X-Operating-System: FreeBSD i386 X-Public-Key: http://www.pittgoth.com/~hiten/pubkey.asc X-URL: http://www.unixdaemons.com/~hiten X-PGP: http://pgp.mit.edu:11371/pks/lookup?search=Hiten+Pandya&op=index cc: freebsd-performance@freebsd.org cc: "Thomas Krause \(Webmatic\)" Subject: Re: apache2 tuning 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: Tue, 06 May 2003 11:39:17 -0000 Sean Chittenden (Mon, May 05, 2003 at 11:37:41AM -0700) wrote: > > I want to build a new webserver (dual xeon with 4 GB RAM). The > > server provides mostly dynamic php-pages. In the ports Makefile > > there are compile option like WITH_THREADS and WITH_MPM (which > > includes WITH_THREADS). These are useful options for a production > > machine? Also, is it useful to compile the kernel with > > ACCEPT_FILTER_DATA and ACCEPT_FILTER_HTTP? And if yes, is apache > > compiled with accept filter by default? > > Having the accept filters turned on in the kernel is a huge win for > Apache. See the tuning(7) man page for other details of interest. As > for which MPM, last I heard UNIX folk were better off using the > pre-fork MPM and not the threaded MPM. This may have changed, but I'd > bet dime to dollar you'll get better stability out of the pre-fork but > a smaller memory footprint out of the threaded MPM. -sc Turning on the Accept Filters is not the only thing you need to do. You will have to explicitly choose the type of accept filter for Apache. Last time I checked, it uses 'dataready', i.e. ACCEPT_FILTER_DATA by default. ACCEPT_FILTER_HTTP does a really good job with Apache and thttpd for that matter. Cheers! -- Hiten From owner-freebsd-performance@FreeBSD.ORG Tue May 6 04:49:26 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 CE48E37B401 for ; Tue, 6 May 2003 04:49:26 -0700 (PDT) Received: from mx02.egartech.com (aloha.egartech.com [62.118.81.133]) by mx1.FreeBSD.org (Postfix) with SMTP id 19EF043F75 for ; Tue, 6 May 2003 04:49:25 -0700 (PDT) (envelope-from temik@egartech.com) Received: (qmail 3682 invoked by uid 85); 6 May 2003 11:49:19 -0000 Received: from temik@egartech.com by mx02.egartech.com with qmail-scanner-1.03 (. Clean. Processed in 0.343934 secs); 06 May 2003 11:49:19 -0000 Received: from unknown (HELO turtle.egar.egartech.com) (192.168.8.4) by 0 with SMTP; 6 May 2003 11:49:18 -0000 content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="koi8-r" Content-Transfer-Encoding: quoted-printable X-MimeOLE: Produced By Microsoft Exchange V6.0.6249.0 Date: Tue, 6 May 2003 15:49:21 +0400 Message-ID: <5235EF9BAE6B7F4CB3735789EEF73B2907425F@turtle.egar.egartech.com> X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: FFS and minfree... Thread-Index: AcMTN9uIDlgnnifCRmCTNuXFofnaKQAizvpw From: "Artem Tepponen" To: Subject: RE: FFS and minfree... 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: Tue, 06 May 2003 11:49:27 -0000 > From: Charles Swiger [mailto:cswiger@mac.com] > Sent: Monday, May 05, 2003 10:47 PM > > Should they be asking 'Why FS design is so flawed > > that it has this [minfree] requirement' instead? >=20 > No. Good answer. Explains everything. > > Is there any working FS for FreeBSD that does not have this=20 > > requirement? >=20 > The FFS doesn't have a requirement that you reserve any free space,=20 > although doing so is a reasonable default for most=20 > circumstances due to the performance advantages. Were those defaults reasonable when disks were small and files usually were in 100k range? I suspect many of my files are a bit larger. More precise question: will I lose anything performance wise if I'll drop minfree to 0% and partition where this happens contains primarily files in 5Mb+ range? > Do you enjoy having to defragment other filesystem types, by=20 > any chance? Background defragmenter/cleaner/whatever running at 3am actually wouldn't bother me too much. But I do enjoy having a few gigs more. There is no such thing as too much storage. Artem From owner-freebsd-performance@FreeBSD.ORG Tue May 6 05:58:19 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 AF17337B401 for ; Tue, 6 May 2003 05:58:19 -0700 (PDT) Received: from otter3.centtech.com (moat3.centtech.com [207.200.51.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id C802643F85 for ; Tue, 6 May 2003 05:58:18 -0700 (PDT) (envelope-from anderson@centtech.com) Received: from centtech.com (electron.centtech.com [204.177.173.173]) by otter3.centtech.com (8.12.3/8.12.3) with ESMTP id h46CwH56092814; Tue, 6 May 2003 07:58:17 -0500 (CDT) (envelope-from anderson@centtech.com) Message-ID: <3EB7B156.2030701@centtech.com> Date: Tue, 06 May 2003 07:57:58 -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: Terry Lambert References: <3801.172.182.185.39.1052141430.squirrel@bluhayz.homeunix.org> <20030505180224.M13866@xy.hartford.edu> <3EB6A981.1030504@centtech.com> <3EB7353F.373470F2@mindspring.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org cc: Kyle Rollin Subject: Re: Real and availible RAM 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: Tue, 06 May 2003 12:58:20 -0000 Terry Lambert wrote: > Eric Anderson wrote: > >>>If you look at the way x86 architecture is designed (and somebody else can >>>feel free to correct me if I'm wrong), but system memory is often used in >>>the caching/shadowing of BIOS. This is where a lot of system memory often >>>goes before the OS is loaded - also, as Rob said, the kernel itself will >>>take up memory before the rest of the OS is booted. >> >>I believe that the BIOS steals the ram prior to the OS booting, so that >>the box does not show that as part of the total ram.. Like: > > > Since the BIOS is where FreeBSD obtains the "real memory" number, > this is pretty unlikely; the BIOS would just lie to FreeBSD about > the number, and be done with it. > [..snipped the blabla..] I think you missed my point - I wasn't saying that FreeBSD is hiding anything, nor was I talking about the "missing 6mb problem" directly. You pretty much repeated my point though - "Since the BIOS is where FreeBSD obtains the "real memory" number" - which is what I said, that the BIOS takes the physical RAM, subtracts what it needs for devices like video cards (that "borrow" from physical main memory), then reports that result to any OS/Apps/etc that ask for it. That is why my system doesn't show the 256Mb of memory it has in it, it shows something like 224Mb (32Mb is set aside for the video card).. Unless I am totally missing your point this morning.. :) Eric -- ------------------------------------------------------------------ Eric Anderson Systems Administrator Centaur Technology Attitudes are contagious, is yours worth catching? ------------------------------------------------------------------ From owner-freebsd-performance@FreeBSD.ORG Tue May 6 05:59:38 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 EAD2E37B401 for ; Tue, 6 May 2003 05:59:38 -0700 (PDT) Received: from otter3.centtech.com (moat3.centtech.com [207.200.51.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 48AD943FD7 for ; Tue, 6 May 2003 05:59:38 -0700 (PDT) (envelope-from anderson@centtech.com) Received: from centtech.com (electron.centtech.com [204.177.173.173]) by otter3.centtech.com (8.12.3/8.12.3) with ESMTP id h46CxZ56092873; Tue, 6 May 2003 07:59:35 -0500 (CDT) (envelope-from anderson@centtech.com) Message-ID: <3EB7B1A4.9090500@centtech.com> Date: Tue, 06 May 2003 07:59:16 -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: Sean Chittenden References: <001901c3137b$cc8d66f0$044ec7c8@AJMNOTEBOOK> <20030506065837.GT94932@perrin.int.nxad.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-performance cc: Alan Jorge Markus Subject: Re: FReebsd 5.0 network performance problem 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: Tue, 06 May 2003 12:59:39 -0000 Sean Chittenden wrote: [..snip..] > Don't use 5.0 for production. If you are going to put anything in > production, use a _very_ recent -CURRENT. Even then, 5.0 isn't as > fast as the -STABLE series at the moment because it's still in an > awkward transition between locking via giant and fine grain locking. > If you'd like to help figure out performance problems for -CURRENT > however, have you tried running any kernel profiling to see where the > kernel is spending it's cycles? -sc What's the appropraite way to run kernel profiling, and how do you read/pull out the results? Eric -- ------------------------------------------------------------------ Eric Anderson Systems Administrator Centaur Technology Attitudes are contagious, is yours worth catching? ------------------------------------------------------------------ From owner-freebsd-performance@FreeBSD.ORG Tue May 6 08:02:17 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 31DF037B401 for ; Tue, 6 May 2003 08:02:17 -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 2B6C643FBD for ; Tue, 6 May 2003 08:02:16 -0700 (PDT) (envelope-from tlambert2@mindspring.com) Received: from pool0297.cvx40-bradley.dialup.earthlink.net ([216.244.43.42] helo=mindspring.com) by puffin.mail.pas.earthlink.net with asmtp (SSLv3:RC4-MD5:128) (Exim 3.33 #1) id 19D3xH-0000EU-00; Tue, 06 May 2003 08:02:08 -0700 Message-ID: <3EB7CE27.99E73B13@mindspring.com> Date: Tue, 06 May 2003 08:00:55 -0700 From: Terry Lambert X-Mailer: Mozilla 4.79 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: Artem Tepponen References: <5235EF9BAE6B7F4CB3735789EEF73B2907425F@turtle.egar.egartech.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-ELNK-Trace: b1a02af9316fbb217a47c185c03b154d40683398e744b8a4a12b462b3a0dc251147e85033f898fe6350badd9bab72f9c350badd9bab72f9c350badd9bab72f9c cc: freebsd-performance@freebsd.org Subject: Re: FFS and minfree... 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: Tue, 06 May 2003 15:02:17 -0000 Artem Tepponen wrote: > > > Should they be asking 'Why FS design is so flawed > > > that it has this [minfree] requirement' instead? > > > > No. > > Good answer. Explains everything. It doesn't, but what you truncated out of the response did. > > > Is there any working FS for FreeBSD that does not have this > > > requirement? > > > > The FFS doesn't have a requirement that you reserve any free space, > > although doing so is a reasonable default for most > > circumstances due to the performance advantages. > > Were those defaults reasonable when disks were small and files > usually were in 100k range? I suspect many of my files are > a bit larger. And the default FS block size is correspondingly larger, now, so it evens out. > More precise question: will I lose anything performance wise > if I'll drop minfree to 0% and partition where this happens > contains primarily files in 5Mb+ range? Yeah. When the disk fills up, files will get corrupted because the minfree needs to be larger than the total number of frags in the FS. By setting minfree too high, you are telling the FS code that it has contiguous (full FS block spaces) available when in fact it doesn't. Be a good citizen: don't lie to the operating system. NB: the "corruption" is more one of unexpected failures than it is one of actual corruption, so don't go getting upset over it: instead, blame your applications for not checking for short writes. > > Do you enjoy having to defragment other filesystem types, by > > any chance? > > Background defragmenter/cleaner/whatever running at 3am > actually wouldn't bother me too much. But I do enjoy having > a few gigs more. There is no such thing as too much storage. "The steady state of disks is full" -- Marshall Kirk McKusick No matter how much you have, you're going to bum your head on a limit, because you will always want more. Better to put the limit some place that won't result in file corruption. -- Terry From owner-freebsd-performance@FreeBSD.ORG Tue May 6 08:05:30 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 0CF2137B401 for ; Tue, 6 May 2003 08:05:30 -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 76F6443FAF for ; Tue, 6 May 2003 08:05:29 -0700 (PDT) (envelope-from tlambert2@mindspring.com) Received: from pool0297.cvx40-bradley.dialup.earthlink.net ([216.244.43.42] helo=mindspring.com) by puffin.mail.pas.earthlink.net with asmtp (SSLv3:RC4-MD5:128) (Exim 3.33 #1) id 19D40R-0000iP-00; Tue, 06 May 2003 08:05:23 -0700 Message-ID: <3EB7CEE9.BBB2A626@mindspring.com> Date: Tue, 06 May 2003 08:04:09 -0700 From: Terry Lambert X-Mailer: Mozilla 4.79 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: Eric Anderson References: <3801.172.182.185.39.1052141430.squirrel@bluhayz.homeunix.org> <20030505180224.M13866@xy.hartford.edu> <3EB6A981.1030504@centtech.com> <3EB7353F.373470F2@mindspring.com> <3EB7B156.2030701@centtech.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-ELNK-Trace: b1a02af9316fbb217a47c185c03b154d40683398e744b8a4a12b462b3a0dc251e70bad61a282b38c350badd9bab72f9c350badd9bab72f9c350badd9bab72f9c cc: freebsd-performance@freebsd.org cc: Kyle Rollin Subject: Re: Real and availible RAM 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: Tue, 06 May 2003 15:05:30 -0000 Eric Anderson wrote: > I think you missed my point - I wasn't saying that FreeBSD is hiding > anything, nor was I talking about the "missing 6mb problem" directly. > You pretty much repeated my point though - "Since the BIOS is where > FreeBSD obtains the "real memory" number" - which is what I said, that > the BIOS takes the physical RAM, subtracts what it needs for devices > like video cards (that "borrow" from physical main memory), then reports > that result to any OS/Apps/etc that ask for it. That is why my system > doesn't show the 256Mb of memory it has in it, it shows something like > 224Mb (32Mb is set aside for the video card).. > > Unless I am totally missing your point this morning.. :) You are. The video memory you are talking about in this case is accounted for by the AGP or a similar video driver. It is not sucked up by the BIOS. At most, the BIOS will "eat" a few K (usually, no more than 4K) that is unknown to the OS. Again, you really need to read the machdep.c code, particularly the "memsize()" function. -- Terry From owner-freebsd-performance@FreeBSD.ORG Tue May 6 08:14:06 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 8FF3A37B401 for ; Tue, 6 May 2003 08:14:06 -0700 (PDT) Received: from otter3.centtech.com (moat3.centtech.com [207.200.51.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id AC94643F75 for ; Tue, 6 May 2003 08:14:05 -0700 (PDT) (envelope-from anderson@centtech.com) Received: from centtech.com (electron.centtech.com [204.177.173.173]) by otter3.centtech.com (8.12.3/8.12.3) with ESMTP id h46FE456015830; Tue, 6 May 2003 10:14:04 -0500 (CDT) (envelope-from anderson@centtech.com) Message-ID: <3EB7D12A.2000104@centtech.com> Date: Tue, 06 May 2003 10:13:46 -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: Terry Lambert References: <3801.172.182.185.39.1052141430.squirrel@bluhayz.homeunix.org> <20030505180224.M13866@xy.hartford.edu> <3EB6A981.1030504@centtech.com> <3EB7353F.373470F2@mindspring.com> <3EB7B156.2030701@centtech.com> <3EB7CEE9.BBB2A626@mindspring.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org cc: Kyle Rollin Subject: Re: Real and availible RAM 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: Tue, 06 May 2003 15:14:06 -0000 Terry Lambert wrote: > Eric Anderson wrote: > >>I think you missed my point - I wasn't saying that FreeBSD is hiding >>anything, nor was I talking about the "missing 6mb problem" directly. >>You pretty much repeated my point though - "Since the BIOS is where >>FreeBSD obtains the "real memory" number" - which is what I said, that >>the BIOS takes the physical RAM, subtracts what it needs for devices >>like video cards (that "borrow" from physical main memory), then reports >>that result to any OS/Apps/etc that ask for it. That is why my system >>doesn't show the 256Mb of memory it has in it, it shows something like >>224Mb (32Mb is set aside for the video card).. >> >>Unless I am totally missing your point this morning.. :) > > > You are. The video memory you are talking about in this case is > accounted for by the AGP or a similar video driver. It is not > sucked up by the BIOS. At most, the BIOS will "eat" a few K > (usually, no more than 4K) that is unknown to the OS. Ahh.. right you are.. thanks for the correction.. > Again, you really need to read the machdep.c code, particularly > the "memsize()" function. Yep - I see now.. thanks for the pointer (again). Eric -- ------------------------------------------------------------------ Eric Anderson Systems Administrator Centaur Technology Attitudes are contagious, is yours worth catching? ------------------------------------------------------------------ From owner-freebsd-performance@FreeBSD.ORG Tue May 6 11:26:02 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 CD82F37B401 for ; Tue, 6 May 2003 11:26:02 -0700 (PDT) Received: from dragon.nuxi.com (trang.nuxi.com [66.93.134.19]) by mx1.FreeBSD.org (Postfix) with ESMTP id 45D5543F3F for ; Tue, 6 May 2003 11:26:02 -0700 (PDT) (envelope-from obrien@NUXI.com) Received: from dragon.nuxi.com (obrien@localhost [127.0.0.1]) by dragon.nuxi.com (8.12.9/8.12.9) with ESMTP id h46IPlm2038175; Tue, 6 May 2003 11:25:51 -0700 (PDT) (envelope-from obrien@dragon.nuxi.com) Received: (from obrien@localhost) by dragon.nuxi.com (8.12.9/8.12.9/Submit) id h46IPkV0038174; Tue, 6 May 2003 11:25:46 -0700 (PDT) Date: Tue, 6 May 2003 11:25:46 -0700 From: "David O'Brien" To: Rob MacGregor Message-ID: <20030506182546.GB37484@dragon.nuxi.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4i X-Operating-System: FreeBSD 5.0-CURRENT Organization: The NUXI BSD Group X-Pgp-Rsa-Fingerprint: B7 4D 3E E9 11 39 5F A3 90 76 5D 69 58 D9 98 7A X-Pgp-Rsa-Keyid: 1024/34F9F9D5 cc: freebsd-performance@freebsd.org Subject: Re: Real and availible RAM X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: freebsd-performance@freebsd.org List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 06 May 2003 18:26:03 -0000 On Mon, May 05, 2003 at 03:44:28PM +0000, Rob MacGregor wrote: > Please DO NOT send me ANY email directly unless it's a privacy issue. > Reply-to mangled to assist those who don't read the above. NOTE: This IS a personal message intended for you to read directly. Please STOP putting a bogus Reply-To: in your emails. MUA's that do proper replies obeying your Reply-To: result in bounces. Set the Reply-To: either the list address, or to dev-null at your site where you shit can it. This is very anti-social and I ask you to stop it or remove yourself from this list. -- -- David (obrien@FreeBSD.org) From owner-freebsd-performance@FreeBSD.ORG Tue May 6 12:23:26 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 1B15E37B401 for ; Tue, 6 May 2003 12:23:26 -0700 (PDT) Received: from bluhayz.homeunix.org (ip68-106-103-50.nv.nv.cox.net [68.106.103.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 337F743F3F for ; Tue, 6 May 2003 12:23:25 -0700 (PDT) (envelope-from dero@bluhayz.homeunix.org) Received: from bluhayz.homeunix.org (localhost [127.0.0.1]) h46JOgbG070772 for ; Tue, 6 May 2003 15:24:47 -0400 (EDT) (envelope-from dero@bluhayz.homeunix.org) Received: from localhost (dero@localhost)h46JOg2h070769 for ; Tue, 6 May 2003 15:24:42 -0400 (EDT) Date: Tue, 6 May 2003 15:24:41 -0400 (EDT) From: agent dero To: freebsd-performance@freebsd.org Message-ID: <20030506152406.P70768@bluhayz.homeunix.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Subject: The magical missing 6 megabytes of mysterious RAM 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: Tue, 06 May 2003 19:23:26 -0000 I don't beleive it can be BIOS sharing the system RAM with other devices, because my box that is in question has a 1MB Graphic card, and sits in the corner. Absolutely no work is done on the box physically. And I think the 1MB graphic card is enough to display BIOS and show the kernel and OS booting. (even if no monitor is attached) I am almost confident that is the kernel loaded into memory. I haven't seen much of a difference in real and availible memory between a custom kernel, and GENERIC. But, before the messages in /var/log/ are createde the FreeBSD boot manager must boot, and then it starts loading the kernel. So I am 99% sure it is the kernel being loaded into the active RAM. I will be able to test this more if I get a box with less RAM, that needs a good compiled kernel. But I think the issue should be put to rest with, the 6 "missing megabytes" being the kernel. (which must always be in the physical RAM) -thanks! --schnip-- > Yes and no. You can't reuse any memory used for page tables, > so that's "pure overhead". You also can't reuse any memory > that has not been marked as swappable. If you allocate memory > in the kernel, it will be permanently subtracted from the "avail" > (it just doesn't give you another message to the effect of any > chanes in "avail", because it would be printf'ing to the console > all the time, if it did), since kernel memory is allocated as > being type-stable. > --schnip to other message-- > I think you missed my point - I wasn't saying that FreeBSD is hiding > anything, nor was I talking about the "missing 6mb problem" directly. > You pretty much repeated my point though - "Since the BIOS is where > FreeBSD obtains the "real memory" number" - which is what I said, that > the BIOS takes the physical RAM, subtracts what it needs for devices > like video cards (that "borrow" from physical main memory), then reports > that result to any OS/Apps/etc that ask for it. That is why my system > doesn't show the 256Mb of memory it has in it, it shows something like > 224Mb (32Mb is set aside for the video card).. > > Unless I am totally missing your point this morning.. :) From owner-freebsd-performance@FreeBSD.ORG Tue May 6 12:48:36 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 2440437B401 for ; Tue, 6 May 2003 12:48:36 -0700 (PDT) Received: from ns1.vagner.com (ns1.vagner.com [65.39.87.179]) by mx1.FreeBSD.org (Postfix) with ESMTP id 34CB643F3F for ; Tue, 6 May 2003 12:48:35 -0700 (PDT) (envelope-from george@vagner.com) Received: (from root@localhost) by ns1.vagner.com (8.12.8/8.12.8) id h46JmWxc099165; Tue, 6 May 2003 12:48:32 -0700 (MST) (envelope-from george@vagner.com) Received: from amd2000 (pcp01602196pcs.manass01.va.comcast.net [68.50.240.22]) by ns1.vagner.com (8.12.9/8.12.8av) with SMTP id h46JmSCC099155; Tue, 6 May 2003 12:48:29 -0700 (MST) (envelope-from george@vagner.com) Message-ID: <008b01c31408$7766de30$16f03244@vagner.com> From: "Laszlo Vagner" To: "agent dero" , References: <20030506152406.P70768@bluhayz.homeunix.org> Date: Tue, 6 May 2003 15:48:21 -0400 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 6.00.2800.1106 X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 X-Spam-Status: No, hits=-12.9 required=4.0 tests=ORIGINAL_MESSAGE,QUOTED_EMAIL_TEXT,REFERENCES version=2.53 X-Spam-Checker-Version: SpamAssassin 2.53 (1.174.2.15-2003-03-30-exp) X-Virus-Scanned: by AMaViS perl-11 X-Mailman-Approved-At: Tue, 06 May 2003 13:16:36 -0700 Subject: Re: The magical missing 6 megabytes of mysterious RAM 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: Tue, 06 May 2003 19:48:36 -0000 typical BIOS chips are about 1 megabyte and there is an option in the bios to copy the flash bios contents into physical ram for more performance on some systems (some can be disabled too). My idea is that everything before the bios and including the kernel is copied to ram. naturally you couldnt run the kernel out of the flash memory or off the hard disk. Additional kernel drivers would take more ram also. ----- Original Message ----- From: "agent dero" To: Sent: Tuesday, May 06, 2003 3:24 PM Subject: The magical missing 6 megabytes of mysterious RAM > I don't beleive it can be BIOS sharing the system RAM with other devices, > because my box that is in question has a 1MB Graphic card, and sits in the > corner. Absolutely no work is done on the box physically. And I think the > 1MB graphic card is enough to display BIOS and show the kernel and OS > booting. (even if no monitor is attached) > > I am almost confident that is the kernel loaded into memory. I haven't > seen much of a difference in real and availible memory between a custom > kernel, and GENERIC. But, before the messages in /var/log/ are createde > the FreeBSD boot manager must boot, and then it starts loading the kernel. > So I am 99% sure it is the kernel being loaded into the active RAM. > I will be able to test this more if I get a box with less RAM, that needs > a good compiled kernel. But I think the issue should be put to rest with, > the 6 "missing megabytes" being the kernel. (which must always be in the > physical RAM) > > -thanks! > --schnip-- > > Yes and no. You can't reuse any memory used for page tables, > > so that's "pure overhead". You also can't reuse any memory > > that has not been marked as swappable. If you allocate memory > > in the kernel, it will be permanently subtracted from the "avail" > > (it just doesn't give you another message to the effect of any > > chanes in "avail", because it would be printf'ing to the console > > all the time, if it did), since kernel memory is allocated as > > being type-stable. > > > --schnip to other message-- > > I think you missed my point - I wasn't saying that FreeBSD is hiding > > anything, nor was I talking about the "missing 6mb problem" directly. > > You pretty much repeated my point though - "Since the BIOS is where > > FreeBSD obtains the "real memory" number" - which is what I said, that > > the BIOS takes the physical RAM, subtracts what it needs for devices > > like video cards (that "borrow" from physical main memory), then reports > > that result to any OS/Apps/etc that ask for it. That is why my system > > doesn't show the 256Mb of memory it has in it, it shows something like > > 224Mb (32Mb is set aside for the video card).. > > > > Unless I am totally missing your point this morning.. :) > > _______________________________________________ > 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 Tue May 6 14:12:24 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 4124637B401 for ; Tue, 6 May 2003 14:12:24 -0700 (PDT) Received: from mailer.neoline.com.br (mailer.neoline.com.br [200.199.96.26]) by mx1.FreeBSD.org (Postfix) with SMTP id 7AF0943F85 for ; Tue, 6 May 2003 14:12:22 -0700 (PDT) (envelope-from freebsd@staff.neoline.com.br) Received: (qmail 32551 invoked from network); 6 May 2003 21:12:20 -0000 Received: from pf04.netwaybbs.com.br (HELO AJMNOTEBOOK) (200.199.78.4) by mailer.neoline.com.br with SMTP; 6 May 2003 21:12:20 -0000 Message-ID: <000b01c31414$7ac94ca0$044ec7c8@AJMNOTEBOOK> From: "Alan Jorge Markus" To: "freebsd-performance" References: <001901c3137b$cc8d66f0$044ec7c8@AJMNOTEBOOK> <20030506065837.GT94932@perrin.int.nxad.com> Date: Tue, 6 May 2003 18:14:28 -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: FReebsd 5.0 network performance problem 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: Tue, 06 May 2003 21:12:24 -0000 > Don't use 5.0 for production. If you are going to put anything in > production, use a _very_ recent -CURRENT. Even then, 5.0 isn't as > fast as the -STABLE series at the moment because it's still in an > awkward transition between locking via giant and fine grain locking. > If you'd like to help figure out performance problems for -CURRENT > however, have you tried running any kernel profiling to see where the > kernel is spending it's cycles? -sc > > -- > Sean Chittenden > Hi Sean, I thought that FreeBSD 5.0-RELEASE was a stable release... so I was thinking in figure out the new feature. I will re-install the system using the 4.8 version. Thanks Alan J Markus From owner-freebsd-performance@FreeBSD.ORG Tue May 6 17:48:36 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 CB7DC37B401 for ; Tue, 6 May 2003 17:48:35 -0700 (PDT) Received: from mailer.neoline.com.br (mailer.neoline.com.br [200.199.96.26]) by mx1.FreeBSD.org (Postfix) with SMTP id 317EB43FB1 for ; Tue, 6 May 2003 17:48:34 -0700 (PDT) (envelope-from freebsd@staff.neoline.com.br) Received: (qmail 20843 invoked from network); 7 May 2003 00:48:32 -0000 Received: from pf04.netwaybbs.com.br (HELO AJMNOTEBOOK) (200.199.78.4) by mailer.neoline.com.br with SMTP; 7 May 2003 00:48:32 -0000 Message-ID: <000b01c31432$aea64b40$044ec7c8@AJMNOTEBOOK> From: "Alan Jorge Markus" To: "freebsd-performance" References: <20030506222617.20639.qmail@web14905.mail.yahoo.com> Date: Tue, 6 May 2003 21:50:40 -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 Subject: Re: FReebsd 5.0 network performance problem 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, 07 May 2003 00:48:37 -0000 Hi Paulo, I will follow you... this was my first production server running the 5. series, and I think you are right, I will wait until the 5.2 release be available. I am using freebsd in the last 2 years , i am in love for this OS : ), but I´m just a baby using it. Your are right, I will follow you, i will wait until 5.2 series at least ... Thank you. Best regards. Alan J Markus > Hey Alan, > > 5.0 is a stable release, but since it is a FreeBSD full of new > technology, you should consider it "experimental". I will not migrate > my servers (4.7/4.8) to the 5. series untill 5.2 at least. I run 5.0 on > a test machine and still it has problems (including performance being > slower than the 4 series). > > best regards, > > Paulo Roberto > > > __________________________________ > Do you Yahoo!? > The New Yahoo! Search - Faster. Easier. Bingo. > http://search.yahoo.com > From owner-freebsd-performance@FreeBSD.ORG Tue May 6 23:06:32 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 4387337B401 for ; Tue, 6 May 2003 23:06:32 -0700 (PDT) Received: from hotmail.com (bay1-f132.bay1.hotmail.com [65.54.245.132]) by mx1.FreeBSD.org (Postfix) with ESMTP id CBA0D43FB1 for ; Tue, 6 May 2003 23:06:31 -0700 (PDT) (envelope-from rob_macgregor@hotmail.com) Received: from mail pickup service by hotmail.com with Microsoft SMTPSVC; Tue, 6 May 2003 23:06:31 -0700 Received: from 80.192.46.45 by by1fd.bay1.hotmail.msn.com with HTTP; Wed, 07 May 2003 06:06:31 GMT X-Originating-IP: [80.192.46.45] X-Originating-Email: [rob_macgregor@hotmail.com] From: "Rob MacGregor" To: freebsd-performance@freebsd.org Date: Wed, 07 May 2003 06:06:31 +0000 Mime-Version: 1.0 Content-Type: text/plain; format=flowed Message-ID: X-OriginalArrivalTime: 07 May 2003 06:06:31.0703 (UTC) FILETIME=[CE4B5270:01C3145E] Subject: Re: Real and availible RAM X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: me@privacy.net List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 07 May 2003 06:06:32 -0000 >From: "David O'Brien" > >NOTE: This IS a personal message intended for you to read directly. > >Please STOP putting a bogus Reply-To: in your emails. MUA's that do >proper replies obeying your Reply-To: result in bounces. Set the >Reply-To: either the list address, or to dev-null at your site where you >shit can it. Well, if I only subscribed to one list I could just set the reply-to appropriately. However I don't. Resetting the reply-to address for every message just isn't possible. I assume it's causing the FreeBSD list software problems? I'll see what I can do - the reply-to for this one *is* a valid address, though it's an auto responder. I'll see if I can find a black hole email address. If anybody knows of one I can use I'd be interested. Sadly hotmail dosn't do one :-) >This is very anti-social and I ask you to stop it or remove yourself from >this list. The reason, should you care (and even if you don't) is that I'm utterly fed up of my volume of email being doubled by people who simply hit "Reply to all" to every email. That and people who seem to assume that because I responded I must be able to solve all their problems. Please DO NOT send me ANY email directly unless it's a privacy issue. Reply-to mangled to assist those who don't read the above. -- Rob | What part of "no" was it you didn't understand? _________________________________________________________________ Worried what your kids see online? Protect them better with MSN 8 http://join.msn.com/?page=features/parental&pgmarket=en-gb&XAPID=186&DI=1059 From owner-freebsd-performance@FreeBSD.ORG Wed May 7 00:10:51 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 53B9D37B401 for ; Wed, 7 May 2003 00:10:51 -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 A934E43FB1 for ; Wed, 7 May 2003 00:10:48 -0700 (PDT) (envelope-from tlambert2@mindspring.com) Received: from pool0555.cvx40-bradley.dialup.earthlink.net ([216.244.44.45] helo=mindspring.com) by puffin.mail.pas.earthlink.net with asmtp (SSLv3:RC4-MD5:128) (Exim 3.33 #1) id 19DJ4a-00006G-00; Wed, 07 May 2003 00:10:41 -0700 Message-ID: <3EB8B0F8.8C40066E@mindspring.com> Date: Wed, 07 May 2003 00:08:40 -0700 From: Terry Lambert X-Mailer: Mozilla 4.79 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: Laszlo Vagner References: <20030506152406.P70768@bluhayz.homeunix.org> <008b01c31408$7766de30$16f03244@vagner.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-ELNK-Trace: b1a02af9316fbb217a47c185c03b154d40683398e744b8a4e85bb7e15dd54c0ca991ec6090960ce2350badd9bab72f9c350badd9bab72f9c350badd9bab72f9c cc: freebsd-performance@freebsd.org cc: agent dero Subject: Re: The magical missing 6 megabytes of mysterious RAM 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, 07 May 2003 07:10:51 -0000 Laszlo Vagner wrote: > typical BIOS chips are about 1 megabyte and there is an option in the > bios to copy the flash bios contents into physical ram for more performance > on some systems (some can be disabled too). > > My idea is that everything before the bios and including the kernel is > copied to > ram. naturally you couldnt run the kernel out of the flash memory or off > the hard disk. Additional kernel drivers would take more ram also. This would only make the memory reported by CMOS different from the memory reported by the BIOS. Since the memory reported by the CMOS is never printed for you to see, that still leaves the kernel for the "real" vs. "avail" difference. BTW, if you read the code, you will see that the BIOS call used gets memory "extents"; the holes in these extents are mapped devices, BIOS, whatever, as reported by the BIOS. It's accounted in "real". To do what you want, you would need to explicitly modify the BIOS on your hardware to *not* count the flash as "mirrored BIOS". Best suggestion is: treat it like a read-only disk, and boot from it normally: let someone else (the loader) do the copy to RAM. -- Terry From owner-freebsd-performance@FreeBSD.ORG Wed May 7 08:14:27 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 E971037B401 for ; Wed, 7 May 2003 08:14:27 -0700 (PDT) Received: from otter3.centtech.com (moat3.centtech.com [207.200.51.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 58E6543FBD for ; Wed, 7 May 2003 08:14:26 -0700 (PDT) (envelope-from anderson@centtech.com) Received: from centtech.com (electron.centtech.com [204.177.173.173]) by otter3.centtech.com (8.12.3/8.12.3) with ESMTP id h47FEJ56039631; Wed, 7 May 2003 10:14:20 -0500 (CDT) (envelope-from anderson@centtech.com) Message-ID: <3EB922B7.2080002@centtech.com> Date: Wed, 07 May 2003 10:13:59 -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: veedee@c7.campus.utcluj.ro References: <3EB67822.3070802@centtech.com> <20030505182756.093fb1c3.sheep.killer@cultdeadsheep.org> <3EB6A0BF.1040803@centtech.com> <20030506042044.GA84589@c7.campus.utcluj.ro> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org cc: Clement Laforet Subject: Re: NAT performance tweaks 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, 07 May 2003 15:14:28 -0000 veedee@c7.campus.utcluj.ro wrote: [..snip..] > If you have a large network behind your NAT server, defining LARGE_NAT in > src/contrib/ipfilter/ip_nat.h and src/sys/contrib/ipfilter/netinet/ip_nat.h > might help. Don't forget to recompile the kernel and ipfilter. > > Strange enough, I used to have huge pings (up to 80ms in a totally switched > gigabit network) after a few hours of utilization before fiddling with > LARGE_NAT. What would you call a "large network"? Is "LARGE_NAT" something that is able to be made into a sysctl tweak (and if so, should it be?)? Eric -- ------------------------------------------------------------------ Eric Anderson Systems Administrator Centaur Technology Attitudes are contagious, is yours worth catching? ------------------------------------------------------------------ From owner-freebsd-performance@FreeBSD.ORG Wed May 7 10:01:58 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 8DAC637B401 for ; Wed, 7 May 2003 10:01:58 -0700 (PDT) Received: from c7.campus.utcluj.ro (c7.campus.utcluj.ro [193.226.6.226]) by mx1.FreeBSD.org (Postfix) with SMTP id 4FD2D43F75 for ; Wed, 7 May 2003 10:01:57 -0700 (PDT) (envelope-from veedee@c7.campus.utcluj.ro) Received: (qmail 13143 invoked by uid 1008); 7 May 2003 17:01:56 -0000 Date: 7 May 2003 20:01:55 +0300 Message-ID: <20030507170155.GA13015@c7.campus.utcluj.ro> From: veedee@c7.campus.utcluj.ro To: "Eric Anderson" References: <3EB67822.3070802@centtech.com> <20030505182756.093fb1c3.sheep.killer@cultdeadsheep.org> <3EB6A0BF.1040803@centtech.com> <20030506042044.GA84589@c7.campus.utcluj.ro> <3EB922B7.2080002@centtech.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3EB922B7.2080002@centtech.com> cc: freebsd-performance@freebsd.org cc: Clement Laforet Subject: Re: NAT performance tweaks 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, 07 May 2003 17:01:58 -0000 On Wed, May 07, 2003 at 10:13:59AM -0500, Eric Anderson wrote: > veedee@c7.campus.utcluj.ro wrote: > [..snip..] > >If you have a large network behind your NAT server, defining LARGE_NAT in > >src/contrib/ipfilter/ip_nat.h and src/sys/contrib/ipfilter/netinet/ip_nat.h > >might help. Don't forget to recompile the kernel and ipfilter. > > > >Strange enough, I used to have huge pings (up to 80ms in a totally switched > >gigabit network) after a few hours of utilization before fiddling with > >LARGE_NAT. > > What would you call a "large network"? according to ip_nat.h... --- #undef LARGE_NAT /* define this if you're setting up a system to NAT * LARGE numbers of networks/hosts - i.e. in the * hundreds or thousands. In such a case, you should * also change the RDR_SIZE and NAT_SIZE below to more * appropriate sizes. The figures below were used for * a setup with 1000-2000 networks to NAT. */ --- > > Is "LARGE_NAT" something that is able to be made into a sysctl tweak > (and if so, should it be?)? Would certainly be nice to make it as a sysctl. It's a pain in the ass to edit it manually everytime I cvsup. :/ > Eric > > -- > ------------------------------------------------------------------ > Eric Anderson Systems Administrator Centaur Technology > Attitudes are contagious, is yours worth catching? > ------------------------------------------------------------------ > -- | Radu Bogdan Rusu | Network Administrator @ campus.utcluj.ro | | cvsup3.ro/www4.ro.freebsd.org maintainer |->5b736c616d215d<-| | Faculty of Automation & Computer Science @ UTCluj , Romania | |-------------------------------------------------------------| From owner-freebsd-performance@FreeBSD.ORG Wed May 7 10:19:09 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 82D0F37B401 for ; Wed, 7 May 2003 10:19:09 -0700 (PDT) Received: from cultdeadsheep.org (charon.cultdeadsheep.org [80.65.226.72]) by mx1.FreeBSD.org (Postfix) with ESMTP id 9159443FAF for ; Wed, 7 May 2003 10:19:07 -0700 (PDT) (envelope-from sheep.killer@cultdeadsheep.org) Received: (qmail 45321 invoked from network); 7 May 2003 17:19:05 -0000 Received: from unknown (HELO lucifer.cultdeadsheep.org) (192.168.0.2) by goofy.cultdeadsheep.org with SMTP; 7 May 2003 17:19:05 -0000 Date: Wed, 7 May 2003 19:19:05 +0200 From: Clement Laforet To: veedee@c7.campus.utcluj.ro Message-Id: <20030507191905.35871f6a.sheep.killer@cultdeadsheep.org> In-Reply-To: <20030507170155.GA13015@c7.campus.utcluj.ro> References: <3EB67822.3070802@centtech.com> <20030505182756.093fb1c3.sheep.killer@cultdeadsheep.org> <3EB6A0BF.1040803@centtech.com> <20030506042044.GA84589@c7.campus.utcluj.ro> <3EB922B7.2080002@centtech.com> <20030507170155.GA13015@c7.campus.utcluj.ro> Organization: tH3 cUlt 0f tH3 d3@d sH33p X-Mailer: Sylpheed version 0.8.11 (GTK+ 1.2.10; i386-portbld-freebsd4.8) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org cc: anderson@centtech.com Subject: Re: NAT performance tweaks 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, 07 May 2003 17:19:09 -0000 > > Would certainly be nice to make it as a sysctl. It's a pain in the ass to > edit it manually everytime I cvsup. :/ > You can surely define NAT_SIZE and others in your make.conf. #ifndef NAT_SIZE # define NAT_SIZE 127 #endif #ifndef RDR_SIZE # define RDR_SIZE 127 #endif #ifndef HOSTMAP_SIZE # define HOSTMAP_SIZE 127 #endif #ifndef NAT_TABLE_SZ # define NAT_TABLE_SZ 127 #endif but don't define LARGE_NAT ie : #ifdef LARGE_NAT #undef NAT_SIZE #undef RDR_SIZE #undef NAT_TABLE_SZ #undef HOSTMAP_SIZE 127 #define NAT_SIZE 2047 #define RDR_SIZE 2047 #define NAT_TABLE_SZ 16383 #define HOSTMAP_SIZE 8191 #endif You should use prime numbers to have a better hash :-) clem From owner-freebsd-performance@FreeBSD.ORG Wed May 7 10:32:49 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 DB29F37B401 for ; Wed, 7 May 2003 10:32:49 -0700 (PDT) Received: from cultdeadsheep.org (charon.cultdeadsheep.org [80.65.226.72]) by mx1.FreeBSD.org (Postfix) with ESMTP id 25D9D43F75 for ; Wed, 7 May 2003 10:32:48 -0700 (PDT) (envelope-from sheep.killer@cultdeadsheep.org) Received: (qmail 45897 invoked from network); 7 May 2003 17:32:46 -0000 Received: from unknown (HELO lucifer.cultdeadsheep.org) (192.168.0.2) by goofy.cultdeadsheep.org with SMTP; 7 May 2003 17:32:46 -0000 Date: Wed, 7 May 2003 19:32:47 +0200 From: Clement Laforet To: veedee@c7.campus.utcluj.ro Message-Id: <20030507193247.6f60584f.sheep.killer@cultdeadsheep.org> In-Reply-To: <20030507170155.GA13015@c7.campus.utcluj.ro> References: <3EB67822.3070802@centtech.com> <20030505182756.093fb1c3.sheep.killer@cultdeadsheep.org> <3EB6A0BF.1040803@centtech.com> <20030506042044.GA84589@c7.campus.utcluj.ro> <3EB922B7.2080002@centtech.com> <20030507170155.GA13015@c7.campus.utcluj.ro> Organization: tH3 cUlt 0f tH3 d3@d sH33p X-Mailer: Sylpheed version 0.8.11 (GTK+ 1.2.10; i386-portbld-freebsd4.8) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org cc: anderson@centtech.com Subject: Re: NAT performance tweaks 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, 07 May 2003 17:32:50 -0000 > > Would certainly be nice to make it as a sysctl. It's a pain in the ass > to edit it manually everytime I cvsup. :/ Using a sysctl variable is a BAD idea for NAT table. If this variable is if it can be) modified, all the NAT table must be reinitialized, because of hash key. You should have a table size which minimizes hash collisions, and then avoids loops and/or hypothetical race conditions. clem From owner-freebsd-performance@FreeBSD.ORG Wed May 7 10:41:02 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 196A337B401 for ; Wed, 7 May 2003 10:41:02 -0700 (PDT) Received: from otter3.centtech.com (moat3.centtech.com [207.200.51.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 38BF243FB1 for ; Wed, 7 May 2003 10:41:01 -0700 (PDT) (envelope-from anderson@centtech.com) Received: from centtech.com (electron.centtech.com [204.177.173.173]) by otter3.centtech.com (8.12.3/8.12.3) with ESMTP id h47Hew56061894; Wed, 7 May 2003 12:40:58 -0500 (CDT) (envelope-from anderson@centtech.com) Message-ID: <3EB94516.5070503@centtech.com> Date: Wed, 07 May 2003 12:40:38 -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: Clement Laforet References: <3EB67822.3070802@centtech.com> <20030505182756.093fb1c3.sheep.killer@cultdeadsheep.org> <3EB6A0BF.1040803@centtech.com> <20030506042044.GA84589@c7.campus.utcluj.ro> <3EB922B7.2080002@centtech.com> <20030507170155.GA13015@c7.campus.utcluj.ro> <20030507193247.6f60584f.sheep.killer@cultdeadsheep.org> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org Subject: Re: NAT performance tweaks 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, 07 May 2003 17:41:02 -0000 Clement Laforet wrote: >>Would certainly be nice to make it as a sysctl. It's a pain in the ass >>to edit it manually everytime I cvsup. :/ > > > Using a sysctl variable is a BAD idea for NAT table. If this variable is > if it can be) modified, all the NAT table must be reinitialized, because > of hash key. You should have a table size which minimizes hash > collisions, and then avoids loops and/or hypothetical race conditions. Well, why not make it a /boot/loader.conf settable sysctl, so while the system is running, it is "read only", and only settable on boot. Is there a flaw in that thinking? Eric -- ------------------------------------------------------------------ Eric Anderson Systems Administrator Centaur Technology Attitudes are contagious, is yours worth catching? ------------------------------------------------------------------ From owner-freebsd-performance@FreeBSD.ORG Wed May 7 10:55:59 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 D923937B401 for ; Wed, 7 May 2003 10:55:59 -0700 (PDT) Received: from cultdeadsheep.org (charon.cultdeadsheep.org [80.65.226.72]) by mx1.FreeBSD.org (Postfix) with ESMTP id E136A43FBF for ; Wed, 7 May 2003 10:55:57 -0700 (PDT) (envelope-from sheep.killer@cultdeadsheep.org) Received: (qmail 46729 invoked from network); 7 May 2003 17:55:56 -0000 Received: from unknown (HELO lucifer.cultdeadsheep.org) (192.168.0.2) by goofy.cultdeadsheep.org with SMTP; 7 May 2003 17:55:56 -0000 Date: Wed, 7 May 2003 19:55:57 +0200 From: Clement Laforet To: Eric Anderson Message-Id: <20030507195557.6554155b.sheep.killer@cultdeadsheep.org> In-Reply-To: <3EB94516.5070503@centtech.com> References: <3EB67822.3070802@centtech.com> <20030505182756.093fb1c3.sheep.killer@cultdeadsheep.org> <3EB6A0BF.1040803@centtech.com> <20030506042044.GA84589@c7.campus.utcluj.ro> <3EB922B7.2080002@centtech.com> <20030507170155.GA13015@c7.campus.utcluj.ro> <20030507193247.6f60584f.sheep.killer@cultdeadsheep.org> <3EB94516.5070503@centtech.com> Organization: tH3 cUlt 0f tH3 d3@d sH33p X-Mailer: Sylpheed version 0.8.11 (GTK+ 1.2.10; i386-portbld-freebsd4.8) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org Subject: Re: NAT performance tweaks 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, 07 May 2003 17:56:00 -0000 > > Well, why not make it a /boot/loader.conf settable sysctl, so while the > system is running, it is "read only", and only settable on boot. > > Is there a flaw in that thinking? It should be great and easily maintainable for sys admins. But I don't know ipfilter maintainer's point of view :) clem From owner-freebsd-performance@FreeBSD.ORG Wed May 7 11:11:12 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 66E7437B401 for ; Wed, 7 May 2003 11:11:12 -0700 (PDT) Received: from perrin.int.nxad.com (internal.ext.nxad.com [69.1.70.251]) by mx1.FreeBSD.org (Postfix) with ESMTP id 4E0BF43F85 for ; Wed, 7 May 2003 11:11:11 -0700 (PDT) (envelope-from sean@perrin.int.nxad.com) Received: by perrin.int.nxad.com (Postfix, from userid 1001) id 6B73520F00; Wed, 7 May 2003 11:11:10 -0700 (PDT) Date: Wed, 7 May 2003 11:11:10 -0700 From: Sean Chittenden To: Clement Laforet Message-ID: <20030507181110.GN49916@perrin.int.nxad.com> References: <3EB67822.3070802@centtech.com> <20030505182756.093fb1c3.sheep.killer@cultdeadsheep.org> <3EB6A0BF.1040803@centtech.com> <20030506042044.GA84589@c7.campus.utcluj.ro> <3EB922B7.2080002@centtech.com> <20030507170155.GA13015@c7.campus.utcluj.ro> <20030507193247.6f60584f.sheep.killer@cultdeadsheep.org> <3EB94516.5070503@centtech.com> <20030507195557.6554155b.sheep.killer@cultdeadsheep.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20030507195557.6554155b.sheep.killer@cultdeadsheep.org> User-Agent: Mutt/1.4i X-PGP-Key: finger seanc@FreeBSD.org X-PGP-Fingerprint: 3849 3760 1AFE 7B17 11A0 83A6 DD99 E31F BC84 B341 X-Web-Homepage: http://sean.chittenden.org/ cc: freebsd-performance@freebsd.org cc: Eric Anderson Subject: Re: NAT performance tweaks 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, 07 May 2003 18:11:12 -0000 > > Well, why not make it a /boot/loader.conf settable sysctl, so > > while the system is running, it is "read only", and only settable > > on boot. > > > > Is there a flaw in that thinking? > > It should be great and easily maintainable for sys admins. But I > don't know ipfilter maintainer's point of view :) At this point, given configuration bits/cold facts have been exchanged regarding what are good settings in ip_nat.h, can we please consider this thread dead? If someone would like to send Darren a patch or an email asking him if'd be interested in making some of the ip_nat.h values into sysctl's, that'd be dandy, however I just assume not see this thread wind on about easier configuration of ipfilter and friends. Performance != making sysadmin life easier. -sc -- Sean Chittenden From owner-freebsd-performance@FreeBSD.ORG Wed May 7 14:45:28 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 31AF337B401 for ; Wed, 7 May 2003 14:45:28 -0700 (PDT) Received: from walter.dfmm.org (walter.dfmm.org [209.151.233.240]) by mx1.FreeBSD.org (Postfix) with ESMTP id 972A743F85 for ; Wed, 7 May 2003 14:45:27 -0700 (PDT) (envelope-from jason@shalott.net) Received: (qmail 48565 invoked by uid 1000); 7 May 2003 21:45:27 -0000 Received: from localhost (sendmail-bs@127.0.0.1) by localhost with SMTP; 7 May 2003 21:45:27 -0000 Date: Wed, 7 May 2003 14:45:26 -0700 (PDT) From: Jason Stone X-X-Sender: To: In-Reply-To: <007b01c314a1$174b0af0$4508a8c0@Beastie> Message-ID: <20030507144226.R4074-100000@walter> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Subject: Re: IPFW Bandwidth throttling? 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, 07 May 2003 21:45:28 -0000 -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 This got mentioned on freebsd-security as a possible tweak to improve the performance of an ipfw firewall using dummynet pipes: > Another thing to maybe try is up the HZ setting in your kernel. Have a > look at the dummynet page. And I've seen other references to increasing the HZ value in order to increase performance with respect to certain network functions. Exactly what does increasing the HZ value do, and what drawbacks might it have? -Jason -------------------------------------------------------------------------- Freud himself was a bit of a cold fish, and one cannot avoid the suspicion that he was insufficiently fondled when he was an infant. -- Ashley Montagu -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (FreeBSD) Comment: See https://private.idealab.com/public/jason/jason.gpg iD8DBQE+uX53swXMWWtptckRAsNKAKDbzy0rLTLtmmPKMZHaBVZFU9788ACguUvI B9x9a9f6JFu7VBPU1QgMK7o= =Ewph -----END PGP SIGNATURE----- From owner-freebsd-performance@FreeBSD.ORG Thu May 8 18:15:13 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 3FADB37B401 for ; Thu, 8 May 2003 18:15:13 -0700 (PDT) Received: from mailer.neoline.com.br (mailer.neoline.com.br [200.199.96.26]) by mx1.FreeBSD.org (Postfix) with SMTP id AF92A43F75 for ; Thu, 8 May 2003 18:15:11 -0700 (PDT) (envelope-from freebsd@staff.neoline.com.br) Received: (qmail 21528 invoked from network); 9 May 2003 01:15:09 -0000 Received: from pf04.netwaybbs.com.br (HELO AJMNOTEBOOK) (200.199.78.4) by mailer.neoline.com.br with SMTP; 9 May 2003 01:15:09 -0000 Message-ID: <000701c315c8$bce3e400$044ec7c8@AJMNOTEBOOK> From: "Alan Jorge Markus" To: "freebsd-performance" Date: Thu, 8 May 2003 22:17:16 -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 Subject: Network best Configuration 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, 09 May 2003 01:15:13 -0000 Hi , I have a freebsd box running squid, and one of the most important things is the network configuration. Just about the network configuration... what kind of improvements I need to add to my system??? I´m running the 4.8 release already with this options in the kernel: options NMBCLUSTERS=32768 and this options in my sysctl.conf: vfs.vmiodirenable=1 net.inet.ip.forwarding=1 kern.ipc.maxsockbuf=2097152 kern.ipc.somaxconn=8192 kern.ipc.max_datalen=512 net.inet.tcp.rfc1323=1 net.inet.tcp.delayed_ack=0 net.inet.tcp.sendspace=65535 net.inet.tcp.recvspace=65535 net.inet.udp.recvspace=65535 net.inet.udp.maxdgram=57344 net.local.stream.recvspace=65535 net.local.stream.sendspace=65535 net.inet.icmp.icmplim=500 kern.ipc.mbuf_wait=64 What more can I do??? Thanks in advance Alan J Markus