Bug 141172 - halt/reboot does not work
Summary: halt/reboot does not work
Status: VERIFIED FIXED
Alias: None
Product: SUSE Linux 10.1
Classification: openSUSE
Component: Xen (show other bugs)
Version: Alpha 4
Hardware: i686 Other
: P5 - None : Normal (vote)
Target Milestone: ---
Assignee: Jan Beulich
QA Contact: E-mail List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2006-01-02 11:00 UTC by peter czanik
Modified: 2006-05-09 09:03 UTC (History)
1 user (show)

See Also:
Found By: Other
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description peter czanik 2006-01-02 11:00:37 UTC
Test was done on a Compaq nx8220 notebook with factory from 30th of December.

When I use the Xen kernel, halt/reboot does not work. They stop at random locations  of the shutdown procedure, sometimes the last line is about shutting down SSH, sometimes it goes a few lines further.

I don't think, as it does not get to the end of shutdown scripts, but probably related to bug #141171
Comment 1 Martin Lasarsch 2006-01-02 15:49:28 UTC
kurt, any ideas? if not please reassign to us ...
Comment 2 Kurt Garloff 2006-01-12 14:44:28 UTC
Hmm, it's a known limitation that poweroff does not work.
(You would need some more ACPI support in the hypervisor for this or a hook
 to call into the dom0 kernel for this.)
But the shutdown procedure should really complete until the kernel displays
halting system. Reboot should work as well.
Comment 3 peter czanik 2006-01-15 08:53:14 UTC
Tested with latest factory (13th of January), it's still there. It also slows down loading of GRUB, probably because the reiserfs filesystem is left in a very dirty state (100+ filesystem fixes on startup).
Comment 4 Kurt Garloff 2006-01-17 11:10:58 UTC
Does this happen always? Regardless of whether you start additional domains
with Xen or not? 
If it only happens with additional domains: Do you use file: (loopback) exports?
The xenblkd seems to leave the loop device busy and thus causes the umount of
the root filesystem to fail.
Comment 5 peter czanik 2006-01-17 11:16:09 UTC
(In reply to comment #4)
> Does this happen always? Regardless of whether you start additional domains
> with Xen or not? 

Yes, always, independent, if other domains are started or not.
Comment 6 Lynn Bendixsen 2006-03-17 16:41:46 UTC
Does this still happen for you with Beta 8?  We are unable to reproduce this in our lab with the latest Beta.
Comment 7 peter czanik 2006-03-17 16:50:47 UTC
It still did not work with beta6, I have yet to install beta8 on this machine. On beta6 it went almost to the end, the only step missing was the actual reboot.

I change to needinfo to myself, to remind me.
Comment 8 peter czanik 2006-03-24 18:58:45 UTC
It's still there in factory as of 22 03 2006. I guess, that the problem is related to that this machine needed reboot=bios for a long time. The machine halts with the last line being: "Please stand by while rebooting machine" (or something similar, but that's how I remember it).
Comment 9 Lynn Bendixsen 2006-03-30 20:26:28 UTC
Jan, this seems similar to another bug you have looked at. Do you have anything you can add here?
Comment 10 Jan Beulich 2006-04-03 08:46:47 UTC
No, the other one is about not being able to reboot - it gets all the way through to shutdown. This here is different, it needs to first be clarified if it crashes in some way. Please provide a complete Xen log (from boot to shutdown) obtained through serial (add com1=<baudrate>,<databits><parity><stopbits> to Xen command line and attach second machine serially to capture output). If that is impossible, boot with kernel parameter no-scroll and report any console (log screen) output.
Comment 11 peter czanik 2006-04-11 11:09:57 UTC
It's an all USB machine, so no serial console :-(
I still suspect, that it's related to #141172 . When I reboot domain0, it gets to the very last message, which is printed before reboot, but the actual reboot does not happen.
Comment 12 Jan Beulich 2006-04-11 12:12:55 UTC
This *is* 141172 - what other bug are you referring to? If indeed you get to the last line, without any hangs or crashes, then this is presumably fixed with the next beta. Please confirm.
Comment 13 peter czanik 2006-04-11 12:16:40 UTC
Ooops. I was thinking about: #141171
Comment 14 peter czanik 2006-04-25 10:44:23 UTC
Works fine with RC2, so it seems to be fixed.
Comment 15 Jan Beulich 2006-04-25 10:49:53 UTC
Thanks.