Dr A V Le Blanc
2006-Mar-16 11:35 UTC
[Xen-users] I/O hangs with Xen 3.0.1 on Dell poweredge
I don''t know if anyone is having a similar problem. I''ve installed Xen 3.0.1 on a Dell PowerEdge 2850 using source downloaded by mercurial on March 14. The machine has an e1000 ethernet card and an LSI megaraid controller which looks like this in the dmesg at boot time: megaraid cmm: 2.20.2.5 (Release Date: Fri Jan 21 00:01:03 EST 2005) megaraid: 2.20.4.5 (Release Date: Thu Feb 03 12:27:22 EST 2005) megaraid: probe new device 0x1028:0x0013:0x1028:0x016d: bus 2:slot 14:func 0 ACPI: PCI Interrupt 0000:02:0e.0[A] -> GSI 46 (level, low) -> IRQ 46 megaraid: fw version:[521S] bios version:[H430] scsi0 : LSI Logic MegaRAID driver I compiled xen after using ''make menuconfig'' to add the MEGARAID_MM and MEGARAID_MAILBOX drivers to the kernel; I do notice that this makes other changes in the .config file. I can boot dom0 and create and start other domains. Unfortunately, whenever I do anything involving much disk access, or so it appears, the kernel appears to hang. I have a DELL DRAC card (for what it''s worth), and I can get in on the serial console and use SysRq to send commands to the hung system, so I know the kernel is still responding, but all ssh sessions active on dom0 or on any other domain hang. The syslog shows no message before the reboots, and there are no particular error messages in xend.log. Has anyone else seen something similar? Is this a known problem, and can it be fixed? Thanks for your comments. -- Owen Dr A V Le Blanc _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Otto Jongerius
2006-Mar-16 12:21 UTC
Re: [Xen-users] I/O hangs with Xen 3.0.1 on Dell poweredge
On Thu, Mar 16, 2006 at 11:35:34AM +0000, Dr A V Le Blanc wrote:> I can boot dom0 and create and start other domains. Unfortunately, > whenever I do anything involving much disk access, or so it appears, > the kernel appears to hang. I have a DELL DRAC card (for what it''s > worth), and I can get in on the serial console and use SysRq to > send commands to the hung system, so I know the kernel is still > responding, but all ssh sessions active on dom0 or on any other > domain hang. The syslog shows no message before the reboots, > and there are no particular error messages in xend.log.Not a solution, but a workaround: I had the same problem and "fixed" it by passing "nousb" to domain0 (which I don''t really need on my servers anyway): title Xen 3.0 NoUSB root (hd0,1) kernel /xen-3.0.1.gz dom0_mem=400000 module /xen-linux-2.6.12.6-xen root=/dev/md0 ro console=tty0 nousb module /xen-modules-2.6.12.6-xen savedefault boot Adding ignorebiostables to the domainU''s should work too. See this post for more info: http://comments.gmane.org/gmane.comp.emulators.xen.devel/12254 Best regards, -- Otto _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
David Ambs
2006-Mar-16 13:50 UTC
Re: [Xen-users] I/O hangs with Xen 3.0.1 on Dell poweredge
On Thu, 16 Mar 2006, Dr A V Le Blanc wrote:> I can boot dom0 and create and start other domains. Unfortunately, > whenever I do anything involving much disk access, or so it appears, > the kernel appears to hang. I have a DELL DRAC card (for what it''s > worth), and I can get in on the serial console and use SysRq to > send commands to the hung system, so I know the kernel is still > responding, but all ssh sessions active on dom0 or on any other > domain hang. The syslog shows no message before the reboots, > and there are no particular error messages in xend.log. > > Has anyone else seen something similar? Is this a known problem, > and can it be fixed? Thanks for your comments.I have this same problem on my box<athlon xp>.... I''m not positive what it is yet.. I''m running Gentoo, so anytime I attempt to upgrade/compile something on Dom0, it causes massive timeouts on dom0/U''s. This started happening after upgrading to Xen 3.0.1/2.6.12.6. I didn''t have the issue in Xen 2.x. I''ve got a test box also<amd64> that is having the issues under 2.6.16. I was sort of wondering if it was the change in the time schedular, or other hardware conflicts. The dom0 kernel is very minimal, and has nothing compiled in but what is required to function as a server. I''ll let ya know if I find anything more out. So far I''ve not seen many people complain of this issue, so I assumed it was my cheap hardware. -dave _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Dr A V Le Blanc
2006-Mar-16 15:47 UTC
[Xen-users] Re: I/O hangs with Xen 3.0.1 on Dell poweredge
On Thu, Mar 16, 2006 at 11:35:34AM +0000, Dr A V Le Blanc wrote:> I can boot dom0 and create and start other domains. Unfortunately, > whenever I do anything involving much disk access, or so it appears, > the kernel appears to hang.... > I can get in on the serial console and use SysRq to > send commands to the hung system, so I know the kernel is still > responding, but all ssh sessions active on dom0 or on any other > domain hang. The syslog shows no message before the reboots, > and there are no particular error messages in xend.log.On Thu 16 Mar 2006 at 13:21:44 +0100, Otto Jongerius <otto.jongerius@atobemobile.com> wrote:> Not a solution, but a workaround: > > I had the same problem and "fixed" it by passing "nousb" to domain0 > (which I don''t really need on my servers anyway): > > title Xen 3.0 NoUSB > root (hd0,1) > kernel /xen-3.0.1.gz dom0_mem=400000 > module /xen-linux-2.6.12.6-xen root=/dev/md0 ro console=tty0 nousb > module /xen-modules-2.6.12.6-xen > savedefault > boot > > Adding ignorebiostables to the domainU''s should work too. See this > post for more info: > > http://comments.gmane.org/gmane.comp.emulators.xen.devel/12254I don''t see the relevance of the note at 12254, but I can confirm that the nousb suggestion definitely appears to work around the problem. I''ve tried hitting the resulting system fairly hard, but it just takes the load without a problem. Many thanks, Otto. -- Owen Dr A V Le Blanc _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Max E Baro
2006-Mar-16 20:22 UTC
RE:[Xen-users] I/O hangs with Xen 3.0.1 on Dell poweredge
Message: 4 Date: Thu, 16 Mar 2006 11:35:34 +0000 From: Dr A V Le Blanc A.V.LeBlanc@man.ac.uk <mailto:A.V.LeBlanc@man.ac.uk> Subject: [Xen-users] I/O hangs with Xen 3.0.1 on Dell poweredge To: xen-users@lists.xensource.com <mailto:xen-users@lists.xensource.com> Message-ID: 20060316113534.GA30223@afs.mcc.ac.uk <mailto:20060316113534.GA30223@afs.mcc.ac.uk> Content-Type: text/plain; charset=us-ascii I don''t know if anyone is having a similar problem. I''ve installed Xen 3.0.1 on a Dell PowerEdge 2850 using source downloaded by mercurial on March 14. The machine has an e1000 ethernet card and an LSI megaraid controller which looks like this in the dmesg at boot time: megaraid cmm: 2.20.2.5 (Release Date: Fri Jan 21 00:01:03 EST 2005) megaraid: 2.20.4.5 (Release Date: Thu Feb 03 12:27:22 EST 2005) megaraid: probe new device 0x1028:0x0013:0x1028:0x016d: bus 2:slot 14:func 0 ACPI: PCI Interrupt 0000:02:0e.0[A] -> GSI 46 (level, low) -> IRQ 46 megaraid: fw version:[521S] bios version:[H430] scsi0 : LSI Logic MegaRAID driver I compiled xen after using ''make menuconfig'' to add the MEGARAID_MM and MEGARAID_MAILBOX drivers to the kernel; I do notice that this makes other changes in the .config file. I can boot dom0 and create and start other domains. Unfortunately, whenever I do anything involving much disk access, or so it appears, the kernel appears to hang. I have a DELL DRAC card (for what it''s worth), and I can get in on the serial console and use SysRq to send commands to the hung system, so I know the kernel is still responding, but all ssh sessions active on dom0 or on any other domain hang. The syslog shows no message before the reboots, and there are no particular error messages in xend.log. Has anyone else seen something similar? Is this a known problem, and can it be fixed? Thanks for your comments. -- Owen Dr A V Le Blanc I painfully went through this and spent over a month trying to figure it out with 2 PE2850''s. One gracious XEN user pointed out that the USB controllers were the source of my problems. I disabled all USB controllers and removed he drivers using ''make menuconfig'' and both servers are pretty much humming along. Not a single hang since then. Now networking is a whole other issue...still need to crack that egg. Max Baro Technical Support Supervisor FACTS Services, Inc. (305) 284 - 7440 meb@factsservices.com _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users