This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


Re: [Xen-devel] domU crashing, trying to determine why

On Tue, 4 Jan 2011 10:54:44 +0000, Ian Campbell <Ian.Campbell@xxxxxxxxxx> wrote:
> On Mon, 2011-01-03 at 16:40 +0000, Konrad Rzeszutek Wilk wrote:
> > > > I've got this setup:
> > > > 
> > > > vcpus=2
> > > > cpus=[1,3]
> > 
> > Is there anything in the Xen hypervisor ring buffer?
> Equally was there anything on the guest console? If it is dying before
> you can look then you can enable logging via XENCONSOLED_TRACE=guest
> in /etc/sysconfig/xencommons or /etc/default/xencommons (depending on
> distro). Logs show up in /var/log/xen/guest/

Because I configured 'on_crash = preserve' I can see the console after
it crashes, but no, there is nothing on the guest console, except for
the login prompt (which of course no longer works).

> If the guest is still running but deadlocked or similar then it might be
> worth sending a few SysRQs to it. To do this press Ctrl-O on the guest
> console and press one of the magic letters, 'h' will give help.
> 't' (tasks) would be a useful start as would 'p' (processor state).

I'll try this on the next crash. It crashed this morning, so I took the
liberty of adding the kernel log parameters so I can get xen hypervisor
ring buffer debugging information for the next crash.

> It might be worth compiling your guest kernel with lockdep support and
> seeing if it complains about anything.

Is this a .config option? If so, what is it?

> Also, I notice you seem to be running the Debian Lenny classic-Xen
> forward port. Are you running the most recent version of that kernel?
> Unfortunately there is not much developer interest in these classic-Xen
> forwards ports these days.

Yes, I am running the up-to-date Debian Lenny kernel. One solution of
course is to upgrade the system to Squeeze to get an overhaul of the
whole subsystem, getting a newer kernel/xen completely... but of course
that is a bigger change, and would not give us any clue as to what is
going on.

> You could also try the -686-bigmem kernel flavour from Lenny which is a
> PVops enabled kernel, or perhaps a Squeeze kernel from backports.org
> (either -686-bigmem or -xen-686, both are pvops kernels). If your
> problem persists with these then there is more chance of it getting
> fixed than with the Lenny -xen-686 flavour.

Do you mean run that kernel in the Dom0, or the DomU? 


Attachment: pgpMBbf9s0lGv.pgp
Description: PGP signature

Xen-devel mailing list