xen "broke" my system

[Carlos Felipe Santacruz]

This is what happened: normally ran 'yum install xen' and that gave 
no problems, but then i rebooted to get to the xen kernel, it resulted 
on black screen (could change consoles with Alt + F2 etc.) so i 
rebooted and tried the regular 3.1.2 kernel with same results. 
uninstalled xen with 'yum history undo' so it would also remove all 
dependencies then rebooted. nothing. 
anyone knows what could have caused this? 
im running f16 x86_64 on an asus g74sx-bbk7(core i7-2630qm, geforce gtx560m, 8GB DDR3 ram)

[Konrad Rzeszutek Wilk]

On Tue, Dec 06, 2011 at 07:14:57AM -0800, Carlos Felipe Santacruz wrote:
>  This is what happened: normally ran 'yum install xen' and
> that gave 
>  no problems, but then i rebooted to get to the xen kernel, it resulted 
>  on black screen (could change consoles with Alt + F2 etc.) so i 
>  rebooted and tried the regular 3.1.2 kernel with same results. 
>  uninstalled xen with 'yum history undo' so it would also remove all 
>  dependencies then rebooted. nothing. 
>  anyone knows what could have caused this?  
No idea. What happens if you boot the Linux kernel with 'debug loglevel=8'
parameters? Does it work?

>  im running f16 x86_64 on an asus g74sx-bbk7(core i7-2630qm, geforce
> gtx560m, 8GB DDR3 ram)  
>  --
>  xen mailing list
>  xen(a)lists.fedoraproject.org
>  https://admin.fedoraproject.org/mailman/listinfo/xen

[Pasi Kärkkäinen]

On Tue, Dec 06, 2011 at 07:14:57AM -0800, Carlos Felipe Santacruz wrote:
>     This is what happened: normally ran 'yum install xen' and
> that gave no
>     problems, but then i rebooted to get to the xen kernel, it resulted on
>     black screen (could change consoles with Alt + F2 etc.) so i rebooted and
>     tried the regular 3.1.2 kernel with same results. uninstalled xen with
>     'yum history undo' so it would also remove all dependencies then rebooted.
>     nothing.
>     anyone knows what could have caused this?
>     im running f16 x86_64 on an asus g74sx-bbk7(core i7-2630qm, geforce
>     gtx560m, 8GB DDR3 ram) 
So read your grub settings from /boot/grub2/ and figure out if you have some wrong or
leftover entries there.

-- Pasi

[M A Young]

On Tue, 6 Dec 2011, Carlos Felipe Santacruz wrote:

>  This is what happened: normally ran 'yum install xen' and
> that gave no
>  problems, but then i rebooted to get to the xen kernel, it resulted on black
>  screen (could change consoles with Alt + F2 etc.) so i rebooted and tried
>  the regular 3.1.2 kernel with same results. uninstalled xen with 'yum
>  history undo' so it would also remove all dependencies then rebooted.
>  nothing.
>  anyone knows what could have caused this?
>  im running f16 x86_64 on an asus g74sx-bbk7(core i7-2630qm, geforce gtx560m,
>  8GB DDR3 ram) 
Xen runs the grub2 update script to give you an option to boot the kernel 
with the hypervisor. This is different from the way Fedora handles grub2 
configuration when you add a kernel, as their method doesn't work with 
xen. However there is a risk that you will lose configuration options if 
the Fedora system hasn't set up grub2 correctly. This did happen while 
Fedora 16 was in beta as grub2 overwrote a configuration file when you 
updated it until they fixed it, but I am not sure if this is still a 
problem.

 	Michael Young

[Marko Ristola]

On 12/06/2011 05:14 PM, Carlos Felipe Santacruz wrote:
>  This is what happened: normally ran 'yum install xen' and
> that gave
>  no problems, but then i rebooted to get to the xen kernel, it resulted
>  on black screen (could change consoles with Alt + F2 etc.) so i
>  rebooted and tried the regular 3.1.2 kernel with same results.
>  uninstalled xen with 'yum history undo' so it would also remove all
>  dependencies then rebooted. nothing.
>  anyone knows what could have caused this?
>  im running f16 x86_64 on an asus g74sx-bbk7(core i7-2630qm, geforce gtx560m, 8GB DDR3
> ram) 
I understood your problem this way:
Booting works: You can switch consoles. Graphical mode or X doesn't work
because you have a black login screen (with or without Xen).
/var/log/messages and /var/log/Xorg.0.log might have some report of the failure.

You might still have a working pre-3.1.2 kernel in Grub2 menu around that works
as a temporary solution.

Are all your RPM packets up to date with "yum upgrade"?
Fedora Bugzilla is a good place to search and report these problems.
NVidia Open source Kernel module developers want to get failure reports to fix
regressions and other problems.
Fedora test kernels are worth to check too: there might already be a fix coming.

Regards,
Marko Ristola

>  --
>  xen mailing list
>  xen(a)lists.fedoraproject.org
>  https://admin.fedoraproject.org/mailman/listinfo/xen

[Carlos Felipe Santacruz]

ok, sorry i didnt update until now.
i traced back the previous problem to the nvidia drivers (or lack thereof) installing the
akmod-nvidia packages solved the "no X" problem.
but now another problem surfaced, at the end of boot, the X display starts, i see the
gnome bg for a few seconds but then i get a "system cant recover, contact a system
administrator" error message, then the only option is a forced shutdown, cant change
consoles or anything.
the stock 3.1.5-2 kernel still works.

the whole /var/log/messages for one boot of the xen kernel:

http://pastebin.com/Eqfgd8gE (im sorry i didnt trim it, kinda new to it and dont know what
is important and what isnt)


________________________________
 From: Marko Ristola <marko.ristola(a)kolumbus.fi&gt;
To: Carlos Felipe Santacruz <csantacruz63(a)yahoo.com&gt; 
Cc: &quot;xen(a)lists.fedoraproject.org&quot; <xen(a)lists.fedoraproject.org&gt; 
Sent: Wednesday, December 7, 2011 3:48 PM
Subject: Re: [Fedora-xen] xen "broke" my system
 
On 12/06/2011 05:14 PM, Carlos Felipe Santacruz wrote:
>  This is what happened: normally ran 'yum install xen' and
> that gave
>  no problems, but then i rebooted to get to the xen kernel, it resulted
>  on black screen (could change consoles with Alt + F2 etc.) so i
>  rebooted and tried the regular 3.1.2 kernel with same results.
>  uninstalled xen with 'yum history undo' so it would also remove all
>  dependencies then rebooted. nothing.
>  anyone knows what could have caused this?
>  im running f16 x86_64 on an asus g74sx-bbk7(core i7-2630qm, geforce gtx560m, 8GB DDR3
> ram) 
I understood your problem this way:
Booting works: You can switch consoles. Graphical mode or X doesn't work
because you have a black login screen (with or without Xen).
/var/log/messages and /var/log/Xorg.0.log might have some report of the failure.

You might still have a working pre-3.1.2 kernel in Grub2 menu around that works
as a temporary solution.

Are all your RPM packets up to date with "yum upgrade"?
Fedora Bugzilla is a good place to search and report these problems.
NVidia Open source Kernel module developers want to get failure reports to fix
regressions and other problems.
Fedora test kernels are worth to check too: there might already be a fix coming.

Regards,
Marko Ristola

>  --
>  xen mailing list
>  xen(a)lists.fedoraproject.org
>  https://admin.fedoraproject.org/mailman/listinfo/xen

[M A Young]

On Sat, 17 Dec 2011, Carlos Felipe Santacruz wrote:

>  ok, sorry i didnt update until now.
>  i traced back the previous problem to the nvidia drivers (or lack thereof)
>  installing the akmod-nvidia packages solved the "no X" problem.
>  but now another problem surfaced, at the end of boot, the X display starts,
>  i see the gnome bg for a few seconds but then i get a "system cant recover,
>  contact a system administrator" error message, then the only option is a
>  forced shutdown, cant change consoles or anything.
>  the stock 3.1.5-2 kernel still works.
>  
>  the whole /var/log/messages for one boot of the xen kernel:
>  http://pastebin.com/Eqfgd8gE (im sorry i didnt trim it, kinda new to it and
>  dont know what is important and what isnt) 
The lines

Dec 17 09:34:43 cfsantacruza00R kernel: [   41.736859] gnome-shell[2632]: 
segfault at 118 ip 0000003edeed5674 sp 00007ffff5dc7f10 error 4 in 
libnvidia-glcore.so.290.10[3eddc00000+1912000]
Dec 17 09:34:44 cfsantacruza00R abrt[2651]: Saved core dump of pid 2632 
(/usr/bin/gnome-shell) to /var/spool/abrt/ccpp-2011-12-17-09:34:43-2632 
(106831872 bytes)

looks suspicious. The core dump it refers to might tell you more.

 	Michael Young

[Carlos Felipe Santacruz]

On Sat, 17 Dec 2011, Michael Young wrote:

> The lines
>
> Dec 17 09:34:43 cfsantacruza00R kernel: [   41.736859] gnome-shell[2632]: segfault at 118
> ip 0000003edeed5674 sp 00007ffff5dc7f10 error 4 in
> libnvidia-glcore.so.290.10[3eddc00000+1912000]
> Dec 17 09:34:44 cfsantacruza00R abrt[2651]: Saved core dump of pid 2632
> (/usr/bin/gnome-shell) to /var/spool/abrt/ccpp-2011-12-17-09:34:43-2632 (106831872 bytes)
>
> looks suspicious. The core dump it refers to might tell you more. 
sorry, noob to coredumps, any info on how to open them?

[Konrad Rzeszutek Wilk]

On Sat, Dec 17, 2011 at 09:30:50AM -0800, Carlos Felipe Santacruz wrote:
>  On Sat, 17 Dec 2011, Michael Young wrote:
>  
>  >The lines
>  >
>  >Dec 17 09:34:43 cfsantacruza00R kernel: [   41.736859] gnome-shell[2632]: segfault at
> 118 ip 0000003edeed5674 sp 00007ffff5dc7f10 error 4 in
> libnvidia-glcore.so.290.10[3eddc00000+1912000]
>  >Dec 17 09:34:44 cfsantacruza00R abrt[2651]: Saved core dump of pid 2632
> (/usr/bin/gnome-shell) to /var/spool/abrt/ccpp-2011-12-17-09:34:43-2632 (106831872 bytes)
>  >
>  >looks suspicious. The core dump it refers to might tell you more.
>  
>  sorry, noob to coredumps, any info on how to open them? 
Use Google. There are tons of instructions.

>  --
>  xen mailing list
>  xen(a)lists.fedoraproject.org
>  https://admin.fedoraproject.org/mailman/listinfo/xen