Xen users - Apr 2008 - Identical Win3K3 domU''s. One panics dom0, other runs fine

Hi there.

I''m new to Xen, but a Linux veteran. I''ve been trying to
install Win2K3
in a domU, with some success but I''ve run into an odd problem. The
first
Win2K3 domU boots, installs, and runs perfectly (as far as I can tell). 
The second boots and starts the install, but can''t finish it. the dom0 
kernel panics. This happens whether the first domU is running or not. 
The message in syslog is:

BUG: unable to handle kernel NULL pointer dereference at virtual address 
0000011c

A bit about my system. Xubuntu 7.10 (feisty) running on an IBM HS21 
blade, with iSCSI SAN. I''m booting from SAN and using separate SAN 
volumes for each domU.

I''m attaching the relevant portion of my syslog along with the two domU
config files. The domU name Cepheus is the one that is working fine. The 
domU named Perseus is the one that causes the panic. Any ideas would be 
greatly appreciated. Like I said, I''m new here, and I''m not
sure I''ve
got everything properly set up, so I won''t get insulted if you ask 
rudimentary questions. :)
-- 
Don Faulkner




_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users

I suppose it doesn''t matter what order the domUs are started in?

And assume that perseus''s vif lines (currently being commented out)
don''t make
a difference?

You might want to report this to xen-devel as it''s clearly a bug of
some sort.
It''s not *necessarily* an upstream Xen bug (as opposed to a bug with
Ubuntu''s
XenLinux, for instance) so you might want to consider upstream bug reporting 
too.

Cheers,
Mark

On Wednesday 02 April 2008, Don Faulkner wrote:
> Hi there.
>
> I''m new to Xen, but a Linux veteran. I''ve been trying to
install Win2K3
> in a domU, with some success but I''ve run into an odd problem. The
first
> Win2K3 domU boots, installs, and runs perfectly (as far as I can tell).
> The second boots and starts the install, but can''t finish it. the
dom0
> kernel panics. This happens whether the first domU is running or not.
> The message in syslog is:
>
> BUG: unable to handle kernel NULL pointer dereference at virtual address
> 0000011c
>
> A bit about my system. Xubuntu 7.10 (feisty) running on an IBM HS21
> blade, with iSCSI SAN. I''m booting from SAN and using separate SAN
> volumes for each domU.
>
> I''m attaching the relevant portion of my syslog along with the two
domU
> config files. The domU name Cepheus is the one that is working fine. The
> domU named Perseus is the one that causes the panic. Any ideas would be
> greatly appreciated. Like I said, I''m new here, and I''m
not sure I''ve
> got everything properly set up, so I won''t get insulted if you ask
> rudimentary questions. :)


-- 
Push Me Pull You - Distributed SCM tool (http://www.cl.cam.ac.uk/~maw48/pmpu/)

_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users

Your assumptions are correct. Order doesn''t matter; perseus always 
causes the panic regardless of whether cepheus is already running or 
not. The vif''s are commented out for debug purposes only.

I''ll report this to both xen-devel and the Ubuntu Kernel Team (since 
they apparently maintain the xen kernel too). Thanks.
-- 
Don

Mark Williamson wrote:
> I suppose it doesn''t matter what order the domUs are started in?
>
> And assume that perseus''s vif lines (currently being commented
out) don''t make
> a difference?
>
> You might want to report this to xen-devel as it''s clearly a bug
of some sort.
> It''s not *necessarily* an upstream Xen bug (as opposed to a bug
with Ubuntu''s
> XenLinux, for instance) so you might want to consider upstream bug
reporting
> too.
>
> Cheers,
> Mark
>
> On Wednesday 02 April 2008, Don Faulkner wrote:
>   
>> Hi there.
>>
>> I''m new to Xen, but a Linux veteran. I''ve been trying
to install Win2K3
>> in a domU, with some success but I''ve run into an odd problem.
The first
>> Win2K3 domU boots, installs, and runs perfectly (as far as I can tell).
>> The second boots and starts the install, but can''t finish it.
the dom0
>> kernel panics. This happens whether the first domU is running or not.
>> The message in syslog is:
>>
>> BUG: unable to handle kernel NULL pointer dereference at virtual
address
>> 0000011c
>>
>> A bit about my system. Xubuntu 7.10 (feisty) running on an IBM HS21
>> blade, with iSCSI SAN. I''m booting from SAN and using separate
SAN
>> volumes for each domU.
>>
>> I''m attaching the relevant portion of my syslog along with the
two domU
>> config files. The domU name Cepheus is the one that is working fine.
The
>> domU named Perseus is the one that causes the panic. Any ideas would be
>> greatly appreciated. Like I said, I''m new here, and
I''m not sure I''ve
>> got everything properly set up, so I won''t get insulted if you
ask
>> rudimentary questions. :)
>>     
>
>
>
>   
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users

I see you''ve reported this on xen-devel now.  That''s probably
the best place
for it...
> Your assumptions are correct. Order doesn''t matter; perseus always
> causes the panic regardless of whether cepheus is already running or
> not. The vif''s are commented out for debug purposes only.
OK.
> I''ll report this to both xen-devel and the Ubuntu Kernel Team
(since
> they apparently maintain the xen kernel too). Thanks.
The Xen project has its on XenLinux based on 2.6.18.  This gets updated with 
new Xen code but obviously it''s quite an old Linux version. 
There''s other
work underway which will eventually make newer Xen-enabled Linuxes the 
official Xen project way of doing things.

Several distros maintain their own Xen kernel with the Xen-ification patches 
forward ported from 2.6.18 to whatever kernel version they''re running. 
As a
result, Ubuntu maintain their own XenLinux patchset, which is different to 
upstream Xen.

To be honest, that situation is a bit of a mess, although it should improve in 
future.  This does introduce the possibility that a bug is in the 
forward-ported patches of the distro, rather than the upstream Xen project 
code, which is why I recommend you talk to your distributor directly as well 
as to xen-devel.

Sorry to not be much help, but I hope that clarifies the situation a bit!

Cheers,
Mark

-- 
Push Me Pull You - Distributed SCM tool (http://www.cl.cam.ac.uk/~maw48/pmpu/)

_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users