1

I have two Virtual Machine Servers, both are similar spec 64 bit machines, same processor, etc. Both are running KVM with LVM based disks. One machine is using CentOS 6 Minimal as the Host OS, the other uses Ubuntu Server 11.04.

I moved the guest machine disks by using dd over ssh, while the guests were off.

I can move Ubuntu server guest machines between the two hosts and they work just fine on either. My Windows guest machine was created on the Ubuntu host and runs fine there, but gets the Blue Screen Error immediately on boot when I try to run it on the CentOS host.

The Error is tiled: IRQL_NOT_LESS_OR_EQUAL and reads:

STOP: 0x0000000A (0x00001016, 0x00000002, 0x00000000, 0x804F8FEC)

The Microsoft support information seems to indicate a hardware problem and suggests things like BIOS updates and removing computer components. Obviously that's not the problem here since the hardware is virtualized.

I tried moving the guest disk again to ensure that data wasn't damaged during the move, but it did not fix the problem.

Has anyone else encountered this problem when trying to move Windows guest machines between KVM hosts? Is there anything that can be done to allow Windows to move smoothly from one machine to another?

My understanding is that with virtualized hardware, the system sees the same hardware devices on any machine and therefor can use the same drivers since the real hardware is hidden from the OS. Are there exceptions to this where drivers could be an issue? And last, is there any way (besides a complete reinstall of Windows) to recover from this problem?

UPDATE:

XML Machine definition:

<domain type='kvm'>
  <name>xpvm2</name>
  <uuid>6ddec00b-9f00-29ad-5197-98264c24cabf</uuid>
  <memory>1048576</memory>
  <currentMemory>1048576</currentMemory>
  <vcpu>1</vcpu>
  <os>
    <type arch='x86_64' machine='rhel6.2.0'>hvm</type>
    <boot dev='hd'/>
  </os>
  <features>
    <acpi/>
    <apic/>
    <pae/>
  </features>
  <clock offset='localtime'/>
  <on_poweroff>destroy</on_poweroff>
  <on_reboot>restart</on_reboot>
  <on_crash>restart</on_crash>
  <devices>
    <emulator>/usr/libexec/qemu-kvm</emulator>
    <disk type='block' device='disk'>
      <driver name='qemu' type='raw'/>
      <source dev='/dev/mapper/mainvg-vm_xpvm2'/>
      <target dev='hda' bus='ide'/>
      <address type='drive' controller='0' bus='0' unit='0'/>
    </disk>
    <controller type='ide' index='0'>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x01' function='0x1'/>
    </controller>
    <interface type='bridge'>
      <mac address='52:54:00:f8:3a:21'/>
      <source bridge='br0'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x03' function='0x0'/>
    </interface>
    <serial type='pty'>
      <target port='0'/>
    </serial>
    <console type='pty'>
      <target type='serial' port='0'/>
    </console>
    <input type='tablet' bus='usb'/>
    <input type='mouse' bus='ps2'/>
    <graphics type='vnc' port='-1' autoport='yes'/>
    <video>
      <model type='vga' vram='9216' heads='1'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/>
    </video>
    <memballoon model='virtio'>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/>
    </memballoon>
  </devices>
</domain>

UPDATE 2

ps -ef | grep qemu (broken into lines for readability)

CentOS Machine:

qemu      9742     1 99 Feb18 ?        22:51:48 /usr/libexec/qemu-kvm
-S -M rhel6.2.0 -enable-kvm
-m 1024
-smp 1,sockets=1,cores=1,threads=1
-name xpvm2
-uuid 6ddec00b-9f00-29ad-5197-98264c24cabf
-nodefconfig
-nodefaults
-chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/xpvm2.monitor,server,nowait
-mon chardev=charmonitor,id=monitor,mode=control
-rtc base=localtime
-> -no-shutdown
-drive file=/dev/mapper/mainvg-vm_xpvm2,if=none,id=drive-ide0-0-0,format=raw
-device ide-drive,bus=ide.0,unit=0,drive=drive-ide0-0-0,id=ide0-0-0,bootindex=1
-> -netdev tap,fd=28,id=hostnet0
-device rtl8139,netdev=hostnet0,id=net0,mac=52:54:00:f8:3a:21,bus=pci.0,addr=0x3
-chardev pty,id=charserial0
-device isa-serial,chardev=charserial0,id=serial0
-usb
-device usb-tablet,id=input0
-vnc 127.0.0.1:5
-vga std
-device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x4
-> root     23564  9518  0 17:38 pts/0    00:00:00 grep qemu

Ubuntu Machine:

    105       1616     1  7 17:46 ?        00:00:14 /usr/bin/kvm -S -M pc-0.14
-enable-kvm
-m 1024
-smp 1,sockets=1,cores=1,threads=1
-name xpvm2
-uuid 6ddec00b-9f00-29ad-5197-98264c24cabf
-nodefconfig
-nodefaults
-chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/xpvm2.monitor,server,nowait
-> -mon chardev=charmonitor,id=monitor,mode=readline
-rtc base=localtime
-> -boot c
-drive file=/dev/mapper/mainvg-vm_xpvm2,if=none,id=drive-ide0-0-0,format=raw
-device ide-drive,bus=ide.0,unit=0,drive=drive-ide0-0-0,id=ide0-0-0
-> -netdev tap,fd=18,id=hostnet0
-device rtl8139,netdev=hostnet0,id=net0,mac=52:54:00:f8:3a:21,bus=pci.0,addr=0x3
-chardev pty,id=charserial0
-device isa-serial,chardev=charserial0,id=serial0
-usb
-device usb-tablet,id=input0
-> -vnc 127.0.0.1:0
-vga std
-device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x4
-> myname      1626  1493  0 17:49 pts/0    00:00:00 grep --color=auto qemu

UPDATE 3:

I have replaced the Ubuntu OS with CentOS on the original machine. Back on the original host but with CentOS now instead, the XP VM still does not boot. This leads me to believe that the problem is related to CentOS vs Ubuntu compatibility, and not compatibility of Host Hardware. Still no idea what to try.

10
  • Centos and Ubuntu have different versions and implementations of KVM, QEMU and libvirt. Some adjustments are probably due. Can you post the xml config of the VM (virsh dumpxml $VMNAME)?
    – dyasny
    Feb 12, 2012 at 21:59
  • Yes, I will post tomorrow when I'm at work and have access. I had to make a few changes to the XML files to get them to define and to boot, but maybe Windows need something more than the Linux guests needed?
    – Nick
    Feb 12, 2012 at 22:41
  • I've added the XML dump above. Sorry for the delay.
    – Nick
    Feb 17, 2012 at 21:50
  • so this is the definition on the centos machine, right? does it look the same on ubuntu? did you generate a new definition, or copy the existing one? besides, what is the full windows error message - it's it's about hal.dll it might be possible to work around
    – dyasny
    Feb 18, 2012 at 12:25
  • That's the def that's currently on the CentOS machine. It was copied from the Ubuntu machine, but I changed the emulator path and machine type to match CentOS's path. I didn't see it mention hal.dll, the error text is just the usual, "if this is the first time... reboot, if you see this again, something is broken." boiler plate. I didn't know it was possible to create new defs (besides writing one by hand). The disk image runs fine if I move it back to its old host.
    – Nick
    Feb 18, 2012 at 18:01

3 Answers 3

1

That stop code, with the first parameter set to something relatively small (in this case, 0x1016) means that some code running in kernel mode, probably a device driver, tried to access virtual memory address 0x1016. This really only happens when that code uses a null pointer (which is invalid) and then adds some offset amount to that, in this case, 0x1016. This is almost certainly a device driver bug, brought about by moving the VM from one environment to one that is similar and related, but different.

I suggest that you move the VM back to its original environment and remove all the paravirtualized drivers. Then try moving it. If it works, re-apply the paravirtualized drivers from the new environment.

6
  • I can try that- I'm not sure if it has para-virtualized drivers though- I thought that Windows XP guests had to be fully-virtualized? Because para-virtualization requires the OS to be aware that it's a VM and Windows XP isn't a "VM aware" OS?
    – Nick
    Feb 20, 2012 at 1:04
  • It's entirely possible to add in some paravirtualization to Windows XP by installing drivers that are meant to work with your virtualization package. I don't know if your VMs have that sort of thing installed. Feb 20, 2012 at 16:55
  • There aren't any para-virtualized drivers in use.
    – Nick
    Feb 25, 2012 at 18:21
  • Then either there's a bug in one of the device drivers that's attempting to interact with the virtual hardware or there's a bug in the virtual hardware emulator that confused the device driver or there's just a large enough difference in behavior between what the emulator does and what real hardware does that the device driver didn't cope with it. With analyzing the dump, you're not going to know more. If you do analyze it, it will probably tell you which driver, and that may allow you to reconfigure the VM in way that works around the problem. Feb 27, 2012 at 17:38
  • Is there a way to get the driver name out of that string of numbers?
    – Nick
    Feb 27, 2012 at 17:50
1

I saw this error every time i shutdown my win7 virtual machine from QEMU/KVM. And finally it was because of the arch parameter. I switch my xml configuration from arch=i686 to arch=x86_64. Then I could gracefully shutdown that virtual machine. Hope this is helpful to you.

0

My guess is the error is comming from the machine type. You have:

<os>
  <type arch='x86_64' machine='rhel6.2.0'>hvm</type>
  <boot dev='hd'/>
</os>

But I don't think ubuntu can run this machine type.

virsh capabilities | grep rhel

This comes up nil on my ubuntu machine. Possibly change this to another windows machine type that is within the ubuntu capabilities list. Run virsh capabilities, look under

<guest>
  <os_type>hvm</os_type>
  <arch name='x86_64'>

Mine are:

  <os>
    <type arch='x86_64' machine='pc-0.12'>hvm</type>
    <boot dev='hd'/>
  </os>
1
  • The VM runs fine on the Ubuntu Host, its the CentOS host that's having problems. I've changed machine to pc, with no change in the error. Although it's still showing -M rhel6.2.0 in the ps output... I think this is because Ubuntu has machine types like pc-0.14 and CentOS only has rhel* machine types. When I moved my other VMS to the CentOS host, I changed pc-0.14 to pc in the XML def and they booted fine. Is it possible to add the pc-0.14 machine type to CentOS?
    – Nick
    Feb 20, 2012 at 1:05

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .