Exported XP Guest reboots in loop on Linux host

[udippel]

Exported my main guest XP using Virtualboxes 'Export Appliance' -> 'Import Appliance' on another host of identical OS (Ubuntu 9.10) and identical Virtualbox (3.1.4). On that new host, it boot-loops forever.
A log is attached, showing the first round until I forcibly power it off. I went through the log, but couldn't really make out why this happens.
Can someone please have a look and help me by pointing out the problem? Thanks,

Uwe

[mpack]

Looking at the log may not show much, since I suspect the VM is working perfectly, and it's XP which is failing. More instructive would be to look at the machine xml files from before and after the move, to identify unintended virtual hardware changes. Look in particular to see whether the IDE controller changed, as that is often the culprit in reports such as these.

Also, once you fix the problem I recommend reconfiguring XP (My Computer) so that it doesn't automatically reset on an error. Also run MergeIDE to make it more tolerant of IDE controller changes.

[heeney.cs]

I have a similar problem in that my XP guest on my ubuntu 9.10 host is always looping when I start my XP VM...it gets about as far as the Windows XP welcome screen. When I was using the 3.1.4 release I disabled the IO APIC feature and that stopped the looping( even though it says that this must be enabled for windows guests ). I just installed the 3.1.6 release and thought I would try and re-enable IO APIC. Now it is looping again but it does not seem to make any difference when I disable the IO APIC feature. Note: The windows guest will start in "Safe" mode without looping.

However, this was not an exported guest. It's the original one.

Any idea what is going on? Note: I have never been able to get my USB to work on my Windows guest. The shared folders work fine.

Thanks,

Chris

[mpack]

"A reboot loop" with an XP guest AFAIK happens because a kernel mode driver faulted, and XP is configured by default to reboot automatically when a BSOD happens. There are many possible causes for a kernel BSOD. The important clue in the OPs case is that the VM was previously working fine, until it was exported and then reimported, which indicates that the BDOD was caused by a VM settings (virtual hardware) change. That is not your problem.

You say that "it says that [IO APIC] must be enabled for Windows guests". I don't know what "it" is, but Windows does not require the IO APIC to be set one way or the other (none of my XP guests have that option enabled). However, once Windows has been installed on a PC (or VM) with the option set one way, you can't easily change it afterwards (at least not without changing the HAL, which is beyond the scope of discussion here). This problem also arises when you P2V a PC which has the option enabled (or not) and fail to set up the target VM in a compatible way.

[udippel]

You say that "it says that [IO APIC] must be enabled for Windows guests". I don't know what "it" is, but Windows does not require the IO APIC to be set one way or the other (none of my XP guests have that option enabled). However, once Windows has been installed on a PC (or VM) with the option set one way, you can't easily change it afterwards (at least not without changing the HAL, which is beyond the scope of discussion here). This problem also arises when you P2V a PC which has the option enabled (or not) and fail to set up the target VM in a compatible way.

Yes. And no. Me, the OP, used the 'export->import' facilities. Should this not leave the settings as they are; respectively carry the settings with it?
Question: If the new host (machine) doesn't support these original settings, would we expect VirtualBox to be silent about this failure; or would we expect it to pop up some info?

Uwe

[mpack]

Should this not leave the settings as they are; respectively carry the settings with it?

Absolutely it should - which is not the same as saying that it does. In fact since you get the BSOD it's pretty obvious that something changed about the hardware seen by the guest. I don't know why you quoted my reply to heeney's IO APIC notion, since I made no suggestion that this applied to you.

However, due to a similar discussion that cropped up elsewhere, I'm reminding that the "something" can be the host CPU - the type of which is of course visible to the guest as well. Certain CPU changes can cause BSODs in the "intelppm.sys" service as the XP guest boots up, and if XP is configured appropriately this in turn causes a reboot. Deleting intelppm.sys (or disabling the "processor" service) cures the problem - after making a backup of the VDI of course.

[Jukser]

Deleting intelppm.sys

Thanks, this helped me.

I run XP on XP (I have many different development enviroments) and exporting guest from my old laptop to new one
caused looped booting.
Actually it booted fine first time, but then I updated windows using windows update which requires boot, then it stared looping.

I restarted guest in safe mode(prevents booting) and renamed intelppm.sys to intelppm_old.sys . (it's under windows/system32/drivers)
Restarted again and now guest works.

[udippel]

Thanks, I have now tried all, from IO APIC to dleting intelppm, but no success.
Finally, I guess it is some IO(-bug?) of Virtualbox? Because I prevented the 'Restart' option in 'My Computer', and now the thing simply hangs, on the Welcome Screen. The Welcome Screen is visible, but does not take keyboard neither mouse input. I rebooted (forcefully) twice, and can confirm that the Welcome Screen simply stands. No BSOD.

Hmm.

[digismack]

Before doing a clean install of Ubuntu 10.04 I exported my Windows XP guest using the "Export Appliance" feature. Afterwards I installed VBox 3.1.6 (same version as before) and imported the appliance. It booted up fine and everything worked fine... until I went to shut it down. Instead of shutting down it rebooted, and then hung and went into a reboot loop as soon as the welcome screen popped up. I deleted the appliance and re-imported it from the backup and the same thing happened. After another deletion and re-import I turned off IO-APIC and it booted fine. I then made a snapshot and shut it down. I booted it again expecting it to freak out again, but it worked fine. I don't recall if I had IO-APIC enabled on the previous install or not, but disabling it in the imported appliance has fixed the issue for me.

[GTheMus]

I had the same issue and it seems as if it has something to do the with the virtualbox guest addititions. When exporting my machine i allready had it installed, and when reimporting the machine i couldn't get it to stop rebooting unless i uninstalled the guest additions. Reinstalled guest additions and now it works fine.
Just for fun :p tried to unistall it before exporting, and as i expected it works perfect after importing.

[joe.l3]

Hi all,

here a possible solution:

http://forums.virtualbox.org/viewtopic. ... 39#p190139

[mpack]

It's a little early to call what you posted there "a solution". You posted a random registry hack, with no clear idea given of the purpose or effect.

[joe.l3]

It's a little early to call what you posted there "a solution". You posted a random registry hack, with no clear idea given of the purpose or effect.

mpack
the regkey 'HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Processor\Start' is used to disable processr.sys driver autostart (C:\Windows\System32\drivers). The driver can be started or stopped also from Services in the Control Panel or by other programs. Sometimes this driver causes a loop reboot automatically of Windows guest operating system. Infact processr.sys driver attempting to perform an unsupported operation inside of the virtual machine (like upgrading the physical processors microcode, changing power state on the physical processor.
In my case this problem occurs when I move a windowsXP guest created on AMD-based host to a Centrino host, here the guest machine continuously restarts and it's accessible only in 'safe mode'. The reason for that is that if processr.sys driver fails once, it not attempts to perform the operation that failed again...
I have disabled processr.sys autostart, setting the regkey value to '4' with no negative side effect. So far, the guest system is not giving me more problems.
My post is not the solution but a possible solution, my environment is still in testing stage.
Currently I'm not running a production environment on guest operative system but only a test machine...so we'll see happen...
joe.l3

[vihack]

i have the same problem but i dont have in services Processors how can i fix it if i delete processr.sys doesnt work

[mpack]

"doesnt work" is not enough information to diagnose a cause.