From owner-freebsd-smp@FreeBSD.ORG Sun Jan 18 00:40:52 2004 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 5FD3216A4CE for ; Sun, 18 Jan 2004 00:40:52 -0800 (PST) Received: from mobrien.com (mobrien.com [66.163.18.200]) by mx1.FreeBSD.org (Postfix) with ESMTP id 4F7F543D4C for ; Sun, 18 Jan 2004 00:40:49 -0800 (PST) (envelope-from mike@mobrien.com) Received: from topgun [66.163.18.193] by mobrien.com with ESMTP (SMTPD32-8.04) id A690FAB2007C; Sun, 18 Jan 2004 03:40:48 -0500 From: "M. J. [Mike] O'Brien" To: Date: Sun, 18 Jan 2004 03:40:08 -0500 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.6604 (9.0.2911.0) X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 In-reply-to: <20040116200139.0FDA916A58F@hub.freebsd.org> Importance: Normal Subject: RE: freebsd-smp Digest, Vol 42, Issue 2 X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: mike@mobrien.com List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 18 Jan 2004 08:40:52 -0000 Hey Jonathon: I don't think the modified GENERIC kernel config will do it for you. Try this fir start at building a custom kernel or at leaast change your GENERIC accordingly: ######################################### machine i386 ident MULTICPU maxusers 0 options SMP # Symmetric MultiProcessor Kernel options APIC_IO # Symmetric (APIC) I/O options IPFIREWALL options IPDIVERT ####options HTT # HyperThreading Technology TURNED OFF!! cpu I686_CPU ######################################### best... Mike Message: 1 Date: Fri, 16 Jan 2004 10:00:30 -0000 From: "Hall J D (ISeLS)" Subject: RE: machdep.hlt_logical_cpus missing in 4.9-Release? To: 'John Baldwin' , "'freebsd-smp@freebsd.org'" Message-ID: Content-Type: text/plain; charset="iso-8859-1" > -----Original Message----- > From: John Baldwin > Sent: 15 January 2004 18:06 > > On Thursday 15 January 2004 05:48 am, Hall J D (ISeLS) wrote: > > I've recently installed 4.9-Release (from the ISOs) on a > twin Xeon system > > and compiled a generic SMP kernel (GENERIC with SMP and > APIC_IO enabled). > > > > The problem I'm having is the system is using the logical > HyperThreading > > CPUs even though I don't want it to. > > > > The errata for 4.9-Release says that the logical CPUs are > prevented from > > executing user processes by default, they are not on my > system. It also > > says you can control this behaviour by change the value of the > > machdep.hlt_logical_cpus sysctl(8) variable. My sysctl > doesn't seem to know > > about this variable. > > > > Well, you can do 'sysctl machdep.hlt_cpus=10 to get the same > effect to turn > off CPU's 1and 3. I am not sure why you are not getting the > sysctls. Are > you familiar with C? I am not familiar with C, but I'm happy to check what I can if you point me in the right direction. The machine I was using for these tests has now gone into production so I'll have to build another before I can test the machdep.hlt_cpus option. I'll do that in the next few days and let you know how it goes. Thanks, Jonathan From owner-freebsd-smp@FreeBSD.ORG Mon Jan 19 07:32:49 2004 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 46F3A16A4CE for ; Mon, 19 Jan 2004 07:32:49 -0800 (PST) Received: from smarthost.enta.net (smarthost.enta.net [195.74.97.231]) by mx1.FreeBSD.org (Postfix) with ESMTP id 200D043D1D for ; Mon, 19 Jan 2004 07:32:47 -0800 (PST) (envelope-from jacs@gnome.co.uk) Received: from smartsmtp.enta.net (smtp.enta.net [195.74.97.230]) by smarthost.enta.net (Postfix) with ESMTP id 6E5D7A1C for ; Mon, 19 Jan 2004 15:35:14 +0000 (GMT) Received: from smtp.enta.net (localhost [127.0.0.1]) by smartsmtp.enta.net (8.12.3/8.12.3) with ESMTP id i0JFjQLC019867 for ; Mon, 19 Jan 2004 15:45:27 GMT (envelope-from jacs@gnome.co.uk) Received: from hawk.gnome.co.uk (81-31-113-153.adsl.entanet.co.uk [81.31.113.153]) by smtp.enta.net (Postfix) with ESMTP id 3EC749EF1F for ; Mon, 19 Jan 2004 15:45:26 +0000 (GMT) Received: from [192.168.123.12] (hawk.gnome.co.uk [192.168.123.12]) by hawk.gnome.co.uk (8.12.10/8.12.9) with ESMTP id i0JFWePs003043 for ; Mon, 19 Jan 2004 15:32:40 GMT (envelope-from jacs@gnome.co.uk) From: Chris Stenton To: freebsd-smp@FreeBSD.ORG Content-Type: text/plain Message-Id: <1074526360.2863.6.camel@hawk.gnome.co.uk> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Mon, 19 Jan 2004 15:32:40 +0000 Content-Transfer-Encoding: 7bit X-MailScanner-Information: Please contact your postmaster for more information X-MailScanner: Found to be clean Subject: newbie SMP questions X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 19 Jan 2004 15:32:49 -0000 Hi, I've just built a dual XEON box running 5.2. Could someone give me some pointers on the following. 1. I can't find a thread on the pro and cons of having HTT enabled. The issue must have been flogged to death somewhere. 2. An old rule of thumb for a single processor machine was to have twice the amount of swap space to physical RAM. Is this still the case for a 4 logical CPU machine? I never understood where that rule of thumb came from:-) Chris From owner-freebsd-smp@FreeBSD.ORG Mon Jan 19 11:38:54 2004 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 0267D16A4CE for ; Mon, 19 Jan 2004 11:38:54 -0800 (PST) Received: from silver.he.iki.fi (helenius.fi [193.64.42.241]) by mx1.FreeBSD.org (Postfix) with ESMTP id 24DD143D54 for ; Mon, 19 Jan 2004 11:38:51 -0800 (PST) (envelope-from pete@he.iki.fi) Received: from he.iki.fi (localhost [127.0.0.1]) by silver.he.iki.fi (8.12.9p2/8.11.4) with ESMTP id i0JJckgr096458; Mon, 19 Jan 2004 21:38:46 +0200 (EET) (envelope-from pete@he.iki.fi) Message-ID: <400C3212.5040907@he.iki.fi> Date: Mon, 19 Jan 2004 21:37:54 +0200 From: Petri Helenius User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031007 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Chris Stenton References: <1074526360.2863.6.camel@hawk.gnome.co.uk> In-Reply-To: <1074526360.2863.6.camel@hawk.gnome.co.uk> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-smp@freebsd.org Subject: Re: newbie SMP questions X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 19 Jan 2004 19:38:54 -0000 Chris Stenton wrote: >Hi, > >I've just built a dual XEON box running 5.2. Could someone give me some >pointers on the following. > > >1. I can't find a thread on the pro and cons of having HTT enabled. >The issue must have been flogged to death somewhere. > > > If you run things that spend considerable time behind Giant, disable HTT. There is also an issue with at least HTT where the system panics on boot asking "AP #n failed panic y/n?". For general application/server use, you probably benefit from HTT. >2. An old rule of thumb for a single processor machine was to have twice >the amount of swap space to physical RAM. Is this still the case for a 4 >logical CPU machine? I never understood where that rule of thumb came >from:-) > > Does >4G swap work alright anyway? Pete From owner-freebsd-smp@FreeBSD.ORG Mon Jan 19 18:26:16 2004 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 674DF16A4CE for ; Mon, 19 Jan 2004 18:26:16 -0800 (PST) Received: from smtp104.mail.sc5.yahoo.com (smtp104.mail.sc5.yahoo.com [66.163.169.223]) by mx1.FreeBSD.org (Postfix) with SMTP id B219C43D41 for ; Mon, 19 Jan 2004 18:26:14 -0800 (PST) (envelope-from kaeru@pd.jaring.my) Received: from unknown (HELO ?219.95.213.228?) (khairil?yusof@219.95.213.228 with plain) by smtp104.mail.sc5.yahoo.com with SMTP; 20 Jan 2004 02:26:13 -0000 From: Khairil Yusof To: Chris Stenton In-Reply-To: <1074526360.2863.6.camel@hawk.gnome.co.uk> References: <1074526360.2863.6.camel@hawk.gnome.co.uk> Content-Type: text/plain Message-Id: <1074565563.77627.59.camel@wolverine.home.net> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.5.2FreeBSD GNOME Team Port Date: Tue, 20 Jan 2004 10:26:03 +0800 Content-Transfer-Encoding: 7bit cc: freebsd-smp@FreeBSD.ORG Subject: Re: newbie SMP questions X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 20 Jan 2004 02:26:16 -0000 On Mon, 2004-01-19 at 15:32 +0000, Chris Stenton wrote: > 1. I can't find a thread on the pro and cons of having HTT enabled. For 5.x releases it should be better (less locking) but in my experience lack of a scheduler that has virtual/physical processor affinity can degrade your performance. This happened often on a client's 4 CPU XEON. 2 cpu intensive processes were often running on the same physical CPU (linux), instead of different ones (ULE+KSE will help with regards for this when it hopefully becomes default on 5.3). You should do simple benchmarks of the apps you intent to run on your server with HT on and off. > 2. An old rule of thumb for a single processor machine was to have twice > the amount of swap space to physical RAM. Is this still the case for a 4 > logical CPU machine? I never understood where that rule of thumb came > from:-) I've read in a few places that the swapping algorithms, are optimised for multiples of 2 or 4. I think tuning(8) mentions it. But I don't think the number of processors affects the swap ration you should have. It's more about how much physical memory, and how heavily loaded memory wise do you think your server is going to be. -- "You keep using that word. I do not think it means what you think it means." FreeBSD 5.2-CURRENT i386 9:52am up 1 day, 9:56, 4 users, load averages: 2.11, 1.50, 1.16 From owner-freebsd-smp@FreeBSD.ORG Tue Jan 20 12:58:21 2004 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 03EE416A4CE for ; Tue, 20 Jan 2004 12:58:21 -0800 (PST) Received: from dirty.research.bell-labs.com (ns1.research.bell-labs.com [204.178.16.6]) by mx1.FreeBSD.org (Postfix) with ESMTP id E3B0943D55 for ; Tue, 20 Jan 2004 12:58:00 -0800 (PST) (envelope-from dong@research.bell-labs.com) Received: from scummy.research.bell-labs.com (H-135-104-2-10.research.bell-labs.com [135.104.2.10])i0KKw0cD018251 for ; Tue, 20 Jan 2004 15:58:00 -0500 (EST) Received: from doom-11.cs.bell-labs.com (doom-11.cs.bell-labs.com [135.104.9.224])i0KKvq0C082032 for ; Tue, 20 Jan 2004 15:57:53 -0500 (EST) Received: from doom-11.cs.bell-labs.com (localhost [127.0.0.1]) i0KKvqvD000830 for ; Tue, 20 Jan 2004 15:57:52 -0500 (EST) (envelope-from dong@doom-11.cs.bell-labs.com) Message-Id: <200401202057.i0KKvqvD000830@doom-11.cs.bell-labs.com> To: freebsd-smp@freebsd.org Date: Tue, 20 Jan 2004 15:57:52 -0500 From: Dong Lin Subject: 5.2R SMP kernel crashes when paging on swap X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 20 Jan 2004 20:58:21 -0000 Does anyone have any suggestions on how to debug the following vm_pager problem in a SMP kernel? Thanks in advance. The kernel crashes when vm_pageout_scan kicks. I have a hard time locating the offending line because there is no panic. The machine just goes blank and reboots. If I add DELAY() at the end of vm_pageout_scan (after Giant unlock), it does not crash but causes the following warning on the console: lock order reversal: 1st vm object(vm object) @ vm/swap_pager.c:1325 2nd swap_pager swhash (swap_pager swhash) @ vm/swap_pager.c:1840 3rd vm object (vm object) @ vm/uma_core.c:873 Stack backtrace: backtrace() witness_lock() _mtx_lock_flags() obj_alloc() slab_zalloc() uma_zone_slab() uma_zalloc_internal() uma_zalloc_arg() swp_pager_meta_build() swap_pager_putpages() default_pager_putpages() vm_pageout_flush() vm_pageout_clean() vm_pageout_scan() vm_pageout() fork_exit() fork_trampoline() This dual processor machine works fine with a non-SMP kernel. The SMP kernel works on a uniprocessor machine as well. my SMP kernel: GENERIC with BOOTP and BOOTP_NFSROOT added to run diskless. my machine configuration: all fs are mounted via md or NFS. The only local storage is a swap disk. nfsserver:/disk4/diskless_root 69481814 61293081 2630188 96% / devfs 1 1 0 100% /dev /dev/md0 7406 1720 5094 25% /etc procfs 4 4 0 100% /proc /dev/md1 31470 242 28712 1% /var nfsserver:/usr 2804906 1406602 1173912 55% /usr % swapinfo Device 1K-blocks Used Avail Capacity /dev/ad0s3b 528144 4600 523544 1% MPTable: Timecounter "i8254" frequency 1193182 Hz quality 0 CPU: Intel Pentium III (864.46-MHz 686-class CPU) Origin = "GenuineIntel" Id = 0x686 Stepping = 6 Features=0x383fbff real memory = 134041600 (127 MB) avail memory = 120590336 (115 MB) FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs cpu0 (BSP): APIC ID: 1 cpu1 (AP): APIC ID: 0 ioapic0: Assuming intbase of 0 ioapic0 irqs 0-23 on motherboard Pentium Pro MTRR support enabled From owner-freebsd-smp@FreeBSD.ORG Thu Jan 22 07:58:40 2004 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 2A22716A4CE for ; Thu, 22 Jan 2004 07:58:40 -0800 (PST) Received: from smtp.uol.com.br (smtpout2.uol.com.br [200.221.11.55]) by mx1.FreeBSD.org (Postfix) with ESMTP id 30F1443D1D for ; Thu, 22 Jan 2004 07:58:38 -0800 (PST) (envelope-from smanioto@netsite.com.br) Received: from smannote (200-158-227-130.dsl.telesp.net.br [200.158.227.130]) by scorpion2.uol.com.br (Postfix) with SMTP id 2111A7E1B; Thu, 22 Jan 2004 13:58:32 -0200 (BRST) Message-ID: <003501c3e100$94faf470$cf00a8c0@smannote> From: "Carlos Eduardo Smanioto" To: , "Dong Lin" References: <200401202057.i0KKvqvD000830@doom-11.cs.bell-labs.com> Date: Thu, 22 Jan 2004 13:58:27 -0200 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 6.00.2720.3000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2727.1300 Subject: Re: 5.2R SMP kernel crashes when paging on swap X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 22 Jan 2004 15:58:40 -0000 I have the same problem in FreeBSD 4.9, 5.1 and 5.2!!! []'s Carlos ----- Original Message ----- From: "Dong Lin" To: Sent: Tuesday, January 20, 2004 6:57 PM Subject: 5.2R SMP kernel crashes when paging on swap > > Does anyone have any suggestions on how to debug the following > vm_pager problem in a SMP kernel? Thanks in advance. > > The kernel crashes when vm_pageout_scan kicks. I have a hard time > locating the offending line because there is no panic. The machine > just goes blank and reboots. > > If I add DELAY() at the end of vm_pageout_scan (after Giant unlock), > it does not crash but causes the following warning on the console: > > lock order reversal: > 1st vm object(vm object) @ vm/swap_pager.c:1325 > 2nd swap_pager swhash (swap_pager swhash) @ vm/swap_pager.c:1840 > 3rd vm object (vm object) @ vm/uma_core.c:873 > Stack backtrace: > backtrace() > witness_lock() > _mtx_lock_flags() > obj_alloc() > slab_zalloc() > uma_zone_slab() > uma_zalloc_internal() > uma_zalloc_arg() > swp_pager_meta_build() > swap_pager_putpages() > default_pager_putpages() > vm_pageout_flush() > vm_pageout_clean() > vm_pageout_scan() > vm_pageout() > fork_exit() > fork_trampoline() > > This dual processor machine works fine with a non-SMP kernel. The SMP > kernel works on a uniprocessor machine as well. > > my SMP kernel: > GENERIC with BOOTP and BOOTP_NFSROOT added to run diskless. > > my machine configuration: > all fs are mounted via md or NFS. The only local storage is a swap > disk. > > nfsserver:/disk4/diskless_root 69481814 61293081 2630188 96% / > devfs 1 1 0 100% /dev > /dev/md0 7406 1720 5094 25% /etc > procfs 4 4 0 100% /proc > /dev/md1 31470 242 28712 1% /var > nfsserver:/usr 2804906 1406602 1173912 55% /usr > > % swapinfo > Device 1K-blocks Used Avail Capacity > /dev/ad0s3b 528144 4600 523544 1% > > MPTable: > Timecounter "i8254" frequency 1193182 Hz quality 0 > CPU: Intel Pentium III (864.46-MHz 686-class CPU) > Origin = "GenuineIntel" Id = 0x686 Stepping = 6 > Features=0x383fbff OV,PAT,PSE36,MMX,FXSR,SSE> > real memory = 134041600 (127 MB) > avail memory = 120590336 (115 MB) > FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs > cpu0 (BSP): APIC ID: 1 > cpu1 (AP): APIC ID: 0 > ioapic0: Assuming intbase of 0 > ioapic0 irqs 0-23 on motherboard > Pentium Pro MTRR support enabled > _______________________________________________ > freebsd-smp@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-smp > To unsubscribe, send any mail to "freebsd-smp-unsubscribe@freebsd.org" From owner-freebsd-smp@FreeBSD.ORG Thu Jan 22 21:24:58 2004 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 1F81816A4CE for ; Thu, 22 Jan 2004 21:24:58 -0800 (PST) Received: from tangerine.metron.com (tangerine.metron.com [192.160.193.14]) by mx1.FreeBSD.org (Postfix) with ESMTP id 8E3C243D45 for ; Thu, 22 Jan 2004 21:24:56 -0800 (PST) (envelope-from mikeobrien@spamcop.net) Received: from c-24-130-251-202.we.client2.attbi.com (c-24-130-251-202.we.client2.attbi.com [24.130.251.202]) by tangerine.metron.com (SMI-8.6/SMI-SVR4) with ESMTP id VAA01868 for ; Thu, 22 Jan 2004 21:24:53 -0800 (PST) X-Envelope-To: Received: from c-24-130-251-202.we.client2.attbi.com (localhost [127.0.0.1]) id i0N5OlPu000268 for ; Thu, 22 Jan 2004 21:24:47 -0800 (PST) (envelope-from mikeobrien@spamcop.net) Message-Id: <200401230524.i0N5OlPu000268@c-24-130-251-202.we.client2.attbi.com> X-Mailer: exmh version 2.6.3 04/04/2003 with nmh-1.0.4 To: freebsd-smp@freebsd.org From: mikeobrien@spamcop.net (Mike O'Brien) Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 22 Jan 2004 21:24:47 -0800 Subject: Dual Xeons on Iwill DP533 motherboard X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 23 Jan 2004 05:24:58 -0000 I run -STABLE (4.9) on an Iwill DP533 motherboard with dual 2.4GHz Xeon processors. I enable hyperthreading, and till now, I've had to put an ugly little hack into the mp_machdep.c file because the Iwill BIOS reported funny APIC ID numbers for the processors. This note, which is more or less for the archives, is to say that I just upgraded the BIOS on this motherboard to the latest release on the Iwill web site (http://www.iwillusa.com). I had not previously done this because no sane person upgrades the BIOS if it's not giving them any problems. To my shock, it turns out I _was_ having BIOS problems. There have been occasional odd hangs of this machine under XP (it's a dual-boot) though FreeBSD has been solid as a rock. The clincher came when the client for Second Life started hanging the machine every five minutes. In desperation I upgraded the BIOS to the June, 2003 (latest as of this message) release, and the problem went away completely. Apparently interrupts were being misvectored or something. That isn't all. FreeBSD would no longer boot until I disabled hyperthreading in the BIOS setup. On a hunch, I backed out my hack, reenabled hyperthreading, and FreeBSD-STABLE now correctly finds the four virtual processor cores without it. Apparently this update fixes the funny APIC numbering. Anyone running an Iwill DP533 motherboard should install this BIOS upgrade, it seems. Mike O'Brien From owner-freebsd-smp@FreeBSD.ORG Sat Jan 24 17:35:13 2004 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 2F0AE16A4CE for ; Sat, 24 Jan 2004 17:35:13 -0800 (PST) Received: from pgh.nepinc.com (pgh.nepinc.com [66.207.129.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id E762443D48 for ; Sat, 24 Jan 2004 17:35:11 -0800 (PST) (envelope-from durham@jcdurham.com) Received: from jimslaptop.home.jcdurham.com (18.gibs5.xdsl.nauticom.net [209.195.184.19]) (authenticated bits=0) by pgh.nepinc.com (8.12.8p2/8.12.8) with ESMTP id i0P1ZQAI000258 for ; Sat, 24 Jan 2004 20:35:26 -0500 (EST) (envelope-from durham@jcdurham.com) From: Jim Durham To: freebsd-smp@freebsd.org Date: Sat, 24 Jan 2004 20:35:05 -0500 User-Agent: KMail/1.5.3 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200401242035.05684.durham@jcdurham.com> Subject: Dell Poweredge 2650 reboots X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 25 Jan 2004 01:35:13 -0000 Hi, I hope this isn't old news, but I can't find exactly this problem in the archives.... I'm getting mysterious reboots on new Dell Poweredge 2650 servers with SMP kernels on dual Xeons. I've been running FreeBSD at our place since 3.2 and I was proud of the fact that in I've only ever had one reboot in almost 5 years and that was a flakey drive. This was running on a motley collection of old single-processor boxes which we are replacing with the Dells. The first Dell went on line in the fall as our mailserver and went for about 3 months and then suddenly rebooted for no reason at all. No error message, no panics, no log entries. It was like you pushed the power button. I was bringing the next box on line as a file server and was halfway through copying about 65 gigs of files to it when it suddenly did the same thing. It also did it one more time when I resumed copying after the reboot. I see lots of mentions of reboots with SMP/FreeBSD, but nothing that fits my case exactly. The really embarrasing thing is that we also have two 2650's running..er..um...Windows 2000 server..there I said it... and they have managed to stay up except for the usual forced reboots to install stuff, etc. BTW, these are RAID-1 boxes with 15K 320mhz SCSI drives and 2 gb ram. They have dual power supplies and they are on UPS power. Anyone else have Dells like this and seen this problem? Thanks, Jim From owner-freebsd-smp@FreeBSD.ORG Sat Jan 24 18:08:49 2004 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 5B5A616A4CE for ; Sat, 24 Jan 2004 18:08:49 -0800 (PST) Received: from mail.fpsn.net (mail.fpsn.net [63.224.69.57]) by mx1.FreeBSD.org (Postfix) with ESMTP id B191743D2F for ; Sat, 24 Jan 2004 18:08:46 -0800 (PST) (envelope-from cfaber@fpsn.net) Received: from fpsn.net (mirc-sucks@unixgr.com [63.224.69.60]) (authenticated bits=0) by mail.fpsn.net (8.12.10/8.12.10) with ESMTP id i0P28eTU010723; Sat, 24 Jan 2004 19:08:41 -0700 (MST) Message-ID: <40132510.8040000@fpsn.net> Date: Sat, 24 Jan 2004 19:08:16 -0700 From: Colin Faber Organization: FPSN.NET Development User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.6b) Gecko/20031205 Thunderbird/0.4 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Jim Durham References: <200401242035.05684.durham@jcdurham.com> In-Reply-To: <200401242035.05684.durham@jcdurham.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-smp@freebsd.org Subject: Re: Dell Poweredge 2650 reboots X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 25 Jan 2004 02:08:49 -0000 Hi Jim, If I recall some of the poweredges were recently recalled by Dell because of some bizarre behavior. You might want to check their site and see if yours is one of the effected ones. Jim Durham wrote: >Hi, > >I hope this isn't old news, but I can't find exactly this problem in the >archives.... > >I'm getting mysterious reboots on new Dell Poweredge 2650 servers with SMP >kernels on dual Xeons. > >I've been running FreeBSD at our place since 3.2 and I was proud of the fact >that in I've only ever had one reboot in almost 5 years and that was a flakey >drive. > >This was running on a motley collection of old single-processor boxes which we >are replacing with the Dells. > >The first Dell went on line in the fall as our mailserver and went for about 3 >months and then suddenly rebooted for no reason at all. No error message, no >panics, no log entries. It was like you pushed the power button. > >I was bringing the next box on line as a file server and was halfway through >copying about 65 gigs of files to it when it suddenly did the same thing. It >also did it one more time when I resumed copying after the reboot. > >I see lots of mentions of reboots with SMP/FreeBSD, but nothing that fits my >case exactly. > >The really embarrasing thing is that we also have two 2650's >running..er..um...Windows 2000 server..there I said it... and they have >managed to stay up except for the usual forced reboots to install stuff, etc. > >BTW, these are RAID-1 boxes with 15K 320mhz SCSI drives and 2 gb ram. They >have dual power supplies and they are on UPS power. > >Anyone else have Dells like this and seen this problem? > >Thanks, > >Jim > > > > > > >_______________________________________________ >freebsd-smp@freebsd.org mailing list >http://lists.freebsd.org/mailman/listinfo/freebsd-smp >To unsubscribe, send any mail to "freebsd-smp-unsubscribe@freebsd.org" > > > > -- Colin Faber FPSN.Net Development staff email: cfaber@fpsn.net