Fedora xen - Apr 2007 - kernel-xen-2.6.20 unusable in fc5/fc6/fc7 (crashing)

Hi lists!

Any plans about kernel-xen package? Current 2.6.20 versions seem to crash
and/or freeze (in dom0) for many people, making them completely unusable. 

kernel-xen-2.6.18/19 packages seem to run fine on same hw. 

One of the bugs: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=234008

Thanks!

-- Pasi

On Wed, Apr 11, 2007 at 08:31:55AM +0800, Adrian Chadd
wrote:
> On Tue, Apr 10, 2007, Ben wrote:
> > Funny, the current 2.6.20 fc5 kernel-xen actually *fixed* my problems 
> > that were introduced about 3 revisions ago....
> 
> Is anyone running 2.6.19-1.2911.6.5.fc6 Xen? I''m going to give
this a try and
> see how stable things become (read: can I bring up a domU or copy some data
> over the network without oopsing..)
I''ve run 2.6.19 and it is definitely alot more stable than the 2.6.20
build of the Xen kernels. You should be able to use it quite successfully.

Regards,
Dan.
-- 
|=- Red Hat, Engineering, Emerging Technologies, Boston.  +1 978 392 2496 -=|
|=-           Perl modules: http://search.cpan.org/~danberr/              -=|
|=-               Projects: http://freshmeat.net/~danielpb/               -=|
|=-  GnuPG: 7D3B9505   F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505  -=|

On Tue, Apr 10, 2007, Ben wrote:
> Funny, the current 2.6.20 fc5 kernel-xen actually *fixed* my problems 
> that were introduced about 3 revisions ago....
Is anyone running 2.6.19-1.2911.6.5.fc6 Xen? I''m going to give this a
try and
see how stable things become (read: can I bring up a domU or copy some data
over the network without oopsing..)




Adrian

Adrian Chadd wrote:
> On Tue, Apr 10, 2007, Ben wrote:
>   
>> Funny, the current 2.6.20 fc5 kernel-xen actually *fixed* my problems 
>> that were introduced about 3 revisions ago....
>>     
>
> Is anyone running 2.6.19-1.2911.6.5.fc6 Xen? I''m going to give
this a try and
> see how stable things become (read: can I bring up a domU or copy some data
> over the network without oopsing..)
>
>
>   
2.6.19-xx works fine for me. I had dom0 freeze with 2.6.20 xen kernels. 
I reverted back to 2.6.19 and things seem to be more stable and work nicely.

- Shashin.
>
> Adrian
>
> --
> Fedora-xen mailing list
> Fedora-xen@redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-xen
>   

-- 
Shashin Shinde
Global Professional Services
Red Hat GmbH
Otto-Hahn-Strasse 20
D-85609 Dornach bei München
Germany

Phone   +49 89 205071 210
Mobile  +49 173 3200830
Fax     +49 89 205071 111
Email:  sshinde@redhat.com

On Wed, Apr 11, 2007, Shashin Shinde wrote:
> 2.6.19-xx works fine for me. I had dom0 freeze with 2.6.20 xen kernels. 
> I reverted back to 2.6.19 and things seem to be more stable and work
nicely.
It seems to be passing my initial burn-in tests fine; save for the 100% CPU
/bin/nash process.


adrian

On 4/11/07, Shashin Shinde <sshinde@redhat.com>
wrote:
>
>
> 2.6.19-xx works fine for me. I had dom0 freeze with 2.6.20 xen kernels.
> I reverted back to 2.6.19 and things seem to be more stable and work
> nicely.


Same case with me, 2.6.20-xx kernel-xen freezes on both dom0 and demU, on
other choice but to fallback to 2.6.19-xx.

Thanks. Askar

Adrian Chadd wrote:
> On Wed, Apr 11, 2007, Shashin Shinde wrote:
> 
>> 2.6.19-xx works fine for me. I had dom0 freeze with 2.6.20 xen kernels.
>> I reverted back to 2.6.19 and things seem to be more stable and work
nicely.
> 
> It seems to be passing my initial burn-in tests fine; save for the 100% CPU
> /bin/nash process.

Perhaps you could run that debian system outsize of xen?




-- 

Cheers
John

-- spambait
1aaaaaaa@coco.merseine.nu  Z1aaaaaaa@coco.merseine.nu

Please do not reply off-list

On Wed, Apr 11, 2007, John Summerfield wrote:
> Adrian Chadd wrote:
> >On Wed, Apr 11, 2007, Shashin Shinde wrote:
> >
> >>2.6.19-xx works fine for me. I had dom0 freeze with 2.6.20 xen
kernels.
> >>I reverted back to 2.6.19 and things seem to be more stable and
work
> >>nicely.
> >
> >It seems to be passing my initial burn-in tests fine; save for the 100%
CPU
> >/bin/nash process.
> 
> 
> Perhaps you could run that debian system outsize of xen?
.. kinda defeats the purpose of running multiple virtual Linux servers
inside Xen, eh? :)




Adrian

On Wed, 11 Apr 2007, Adrian Chadd wrote:
> > >>2.6.19-xx works fine for me. I had dom0 freeze with 2.6.20 xen
kernels.
> > >>I reverted back to 2.6.19 and things seem to be more stable
and work
> > >>nicely.
> > >
> > >It seems to be passing my initial burn-in tests fine; save for the
100% CPU
> > >/bin/nash process.
> >
> >
> > Perhaps you could run that debian system outsize of xen?
>
> .. kinda defeats the purpose of running multiple virtual Linux servers
> inside Xen, eh? :)
For older OS''es, you run into this 100% nash/hotplug issue. As a work
around i use a "killall" in the rc.sysvinit script. I''m not
sure what
the real bug is, but I guess nash/hotplug are making some invalid assumptions
about the capability of the kernel, and it should just abort instead
of getting into this infinite loop.

Since the fix is easy, I never delved into what it is actually doing. If
there is an interest in fixing this bug, I could investigate.

Paul

On Thu, 12 Apr 2007, Adrian Chadd wrote:
> On Wed, Apr 11, 2007, Paul Wouters wrote:
>
> > For older OS''es, you run into this 100% nash/hotplug issue.
As a work
> > around i use a "killall" in the rc.sysvinit script.
I''m not sure what
> > the real bug is, but I guess nash/hotplug are making some invalid
assumptions
> > about the capability of the kernel, and it should just abort instead
> > of getting into this infinite loop.
> >
> > Since the fix is easy, I never delved into what it is actually doing.
If
> > there is an interest in fixing this bug, I could investigate.
>
> Weirdly its with a fresh debootstrap''ed Debian etch (so not
"old") Linux
> install under FC6; etch under FC5 with an older FC5 2.6.18 kernel
doesn''t
> trigger the nash CPU loop.
Maybe Debian is still using hotplug instead of udev? Try apt-get install udev

Paul

On Wed, Apr 11, 2007, Paul Wouters wrote:
> For older OS''es, you run into this 100% nash/hotplug issue. As a
work
> around i use a "killall" in the rc.sysvinit script. I''m
not sure what
> the real bug is, but I guess nash/hotplug are making some invalid
assumptions
> about the capability of the kernel, and it should just abort instead
> of getting into this infinite loop.
> 
> Since the fix is easy, I never delved into what it is actually doing. If
> there is an interest in fixing this bug, I could investigate.
Weirdly its with a fresh debootstrap''ed Debian etch (so not
"old") Linux
install under FC6; etch under FC5 with an older FC5 2.6.18 kernel
doesn''t
trigger the nash CPU loop.

I''m running the Debian etch domU with the FC6 Xen kernel (and am
putting
the modules into the domU manually so things like xennet and iptables work
for me) with the 2.6.19-1.2911.6.5.fc6xen kernel and a rebuild mkinitrd
(ie, with xenblk.)

I''d be interested in finding out more about what its actually doing
if only to understand things a little more.



Adrian

On Wed, Apr 11, 2007, Paul Wouters wrote:
> > Weirdly its with a fresh debootstrap''ed Debian etch (so not
"old") Linux
> > install under FC6; etch under FC5 with an older FC5 2.6.18 kernel
doesn''t
> > trigger the nash CPU loop.
> 
> Maybe Debian is still using hotplug instead of udev? Try apt-get install
udev
Already thought about that (and I''ve been bitten by the lack of udev in
default
installs of earlier Debians):

test-1:~# dpkg -l | grep udev
ii  udev                   0.105-4                         /dev/ and hotplug
management daemon
test-1:~# dpkg -l | grep hotplug
ii  udev                   0.105-4                         /dev/ and hotplug
management daemon
test-1:~# mount
/dev/sda1 on / type ext3 (rw,errors=remount-ro)
tmpfs on /lib/init/rw type tmpfs (rw,nosuid,mode=0755)
proc on /proc type proc (rw,noexec,nosuid,nodev)
sysfs on /sys type sysfs (rw,noexec,nosuid,nodev)
procbususb on /proc/bus/usb type usbfs (rw)
udev on /dev type tmpfs (rw,mode=0755)
tmpfs on /dev/shm type tmpfs (rw,nosuid,nodev)
devpts on /dev/pts type devpts (rw,noexec,nosuid,gid=5,mode=620)





Adrian

Hi!

2.6.20 also keeps crashing dom0 and domU here.
Went back to 2.6.19, now it works again.

chris


-- 

Christian Reiter	|||	c.reiter@gmx.net