2:43 a.m.
I looked at some images in /var/lib/xen/images/ and since they have
partition tables and look, to me, just like disks, I copied part of my
hard disk to /var/lib/xen/images/WindowsXP.img
The reason is that I want to try the Windows XP that came preinstalled,
and copying /dev/sda for a length I calculated includes the entire
/dev/sda1 partition seems a likely way to test it.
These are the commands:
dd if=/dev/sda of=/var/lib/xen/images/WindowsXP.img count=76597082
dd of=/var/lib/xen/images/WindowsXP.img seek=320072933376 bs=1 count=0
the second makes the disk image seem to be the proper size.
I tried to start it like so:
qemu-kvm -boot c -snapshot -m 256 -smp 2 -std-vga \
/var/lib/xen/images/WindowsXP.img
and it choked:
[root@potoroo ~]# qemu-kvm -boot c -snapshot -m 256 -smp 2 -std-vga
/var/lib/xen/images/WindowsXP.img
vram_pointer=0x2aaac0b5d000
exception 0 (0)
rax 0000000000000022 rbx 0000000000007000 rcx 000000000064e165 rdx
0000000000000100
rsi 00000000ffff7c05 rdi 000000000008ffea rsp 0000000000001ffe rbp
0000000000000000
r8 0000000000000000 r9 0000000000000000 r10 0000000000000000 r11
0000000000000000
r12 0000000000000000 r13 0000000000000000 r14 0000000000000000 r15
0000000000000000
rip 000000000000fce2 rflags 00033206
cs 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
ds 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
es 0800 (00008000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
ss 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
fs 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
gs 0000 (00000000/0000ffff p 1 dpl 3 db 0 s 1 type 3 l 0 g 0 avl 0)
tr 0000 (10850000/00002088 p 1 dpl 0 db 0 s 0 type b l 0 g 0 avl 0)
ldt 0000 (00000000/0000ffff p 1 dpl 0 db 0 s 0 type 2 l 0 g 0 avl 0)
gdt fad00/30
idt 0/3ff
cr0 60000010 cr2 0 cr3 0 cr4 0 cr8 0 efer 0
code: d5 03 e4 fc 24 44 d5 03 3e 00 d4 03 00 c0 f8 01 00 00 d5 03 --> f6
fc 24 44 d5 03 3f 00 d4 03 00 c0 f8 01 00 00 d5 03 08 fd 24 44 d5 03 40
00 d4 03 00 c0
Aborted
[root@potoroo ~]#
Here is the source disk:
Disk /dev/sda: 320.0 GB, 320072933376 bytes
255 heads, 63 sectors/track, 38913 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0xd42ad42a
Device Boot Start End Blocks Id System
/dev/sda1 * 1 8669 69633711 7 HPFS/NTFS
/dev/sda2 8671 9715 8393962+ 7 HPFS/NTFS
/dev/sda3 9716 9728 104422+ 72 Unknown
/dev/sda4 9729 38913 234428512+ 5 Extended
/dev/sda5 9729 9741 104391 83 Linux
/dev/sda6 9742 38913 234324058+ 8e Linux LVM
[root@potoroo ~]#
Partitions 2 and up are unimportant.
fdisk on the disk image reports:
Command (m for help): p
Disk /var/lib/xen/images/WindowsXP.img: 0 MB, 0 bytes
255 heads, 63 sectors/track, 38913 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0xd42ad42a
Device Boot Start End
Blocks Id System
/var/lib/xen/images/WindowsXP.img1 * 1 8669
69633711 7 HPFS/NTFS
/var/lib/xen/images/WindowsXP.img2 8671 9715
8393962+ 7 HPFS/NTFS
/var/lib/xen/images/WindowsXP.img3 9716 9728
104422+ 72 Unknown
/var/lib/xen/images/WindowsXP.img4 9729 38913
234428512+ 5 Extended
Command (m for help): q
but I had to specify the number of cylinder, 38913 but qemu-kvm rejects
that nonsense.
Is there a better way to do this? I really want to test running this
Windows under Linux before allowing it to use the real disk.
--
Cheers
John
-- spambait
1aaaaaaa(a)coco.merseine.nu Z1aaaaaaa(a)coco.merseine.nu
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375
You cannot reply off-list:-)
4:48 a.m.
John Summerfield wrote:
I looked at some images in /var/lib/xen/images/ and since they have
partition tables and look, to me, just like disks, I copied part of my
hard disk to /var/lib/xen/images/WindowsXP.img
The reason is that I want to try the Windows XP that came preinstalled,
and copying /dev/sda for a length I calculated includes the entire
/dev/sda1 partition seems a likely way to test it.
These are the commands:
dd if=/dev/sda of=/var/lib/xen/images/WindowsXP.img count=76597082
dd of=/var/lib/xen/images/WindowsXP.img seek=320072933376 bs=1 count=0
I don't understand this second command. What's the input? Are you
typing something on stdin (eg. ^D)?
Rich.
--
Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/
Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod
Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in
England and Wales under Company Registration No. 03798903
2:32 a.m.
Richard W.M. Jones wrote:
John Summerfield wrote:
>
> I looked at some images in /var/lib/xen/images/ and since they have
> partition tables and look, to me, just like disks, I copied part of my
> hard disk to /var/lib/xen/images/WindowsXP.img
>
> The reason is that I want to try the Windows XP that came
> preinstalled, and copying /dev/sda for a length I calculated includes
> the entire /dev/sda1 partition seems a likely way to test it.
> These are the commands:
> dd if=/dev/sda of=/var/lib/xen/images/WindowsXP.img count=76597082
> dd of=/var/lib/xen/images/WindowsXP.img seek=320072933376 bs=1 count=0
I don't understand this second command. What's the input? Are you
typing something on stdin (eg. ^D)?
count=0 -- doesn't read anything.
Its purpose is to make a 320 Gbyte file, to match the source disk in
size. It becomes a sparse file, effectively with lots of zeroes between
the end of data copied with the first command.
I've since edited the partition table with GNU Parted, it now looks like
this:
[root@potoroo ~]# parted /var/lib/xen/images/WindowsXP.img
GNU Parted 1.8.6
Using /var/lib/xen/images/WindowsXP.img
Welcome to GNU Parted! Type 'help' to view a list of commands.
(parted) print
Model: (file)
Disk /var/lib/xen/images/WindowsXP.img: 320GB
Sector size (logical/physical): 512B/512B
Partition Table: msdos
Number Start End Size Type File system Flags
1 32.3kB 71.3GB 71.3GB primary ntfs boot
(parted) q
[root@potoroo ~]#
It approximates to how I've copied Windows systems around on real
hardware from time to time. If I were to copy that file to a real disk
and boot in real hardware, I would fully expect it to work.
how embarrassing. I needed to copy about 80 Gbytes, and only copied half
that.
Worse, the boot sector has grub in it.
The good news is that the system now boots grub sans menu. I typed in
the necessary information to boot windows:
chainloader (hd0,0)+1
boot
and windows boots all the way to a BSOD and the stuff on the screen
recommends "chkdsk /f."
Hmm. Type to find a rescue system.
--
Cheers
John
-- spambait
1aaaaaaa(a)coco.merseine.nu Z1aaaaaaa(a)coco.merseine.nu
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375
You cannot reply off-list:-)
2:58 a.m.
John Summerfield wrote:
Richard W.M. Jones wrote:
> John Summerfield wrote:
The good news is that the system now boots grub sans menu. I typed in
the necessary information to boot windows:
chainloader (hd0,0)+1
boot
and windows boots all the way to a BSOD and the stuff on the screen
recommends "chkdsk /f."
Hmm. Type to find a rescue system.
fwiw same happens when I boot directly from the partition. Ideas welcome.
--
Cheers
John
-- spambait
1aaaaaaa(a)coco.merseine.nu Z1aaaaaaa(a)coco.merseine.nu
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375
You cannot reply off-list:-)
4:33 a.m.
John Summerfield wrote:
John Summerfield wrote:
> Richard W.M. Jones wrote:
>> John Summerfield wrote:
>
> The good news is that the system now boots grub sans menu. I typed in
> the necessary information to boot windows:
> chainloader (hd0,0)+1
> boot
>
> and windows boots all the way to a BSOD and the stuff on the screen
> recommends "chkdsk /f."
>
> Hmm. Type to find a rescue system.
fwiw same happens when I boot directly from the partition. Ideas welcome.
Can you capture a screenshot of the BSOD? Maybe some info there.
--
error compiling committee.c: too many arguments to function
6:34 p.m.
Avi Kivity wrote:
John Summerfield wrote:
> John Summerfield wrote:
>> Richard W.M. Jones wrote:
>>> John Summerfield wrote:
>
>>
>> The good news is that the system now boots grub sans menu. I typed in
>> the necessary information to boot windows:
>> chainloader (hd0,0)+1
>> boot
>>
>> and windows boots all the way to a BSOD and the stuff on the screen
>> recommends "chkdsk /f."
>>
>> Hmm. Type to find a rescue system.
>
> fwiw same happens when I boot directly from the partition. Ideas welcome.
>
>
>
Can you capture a screenshot of the BSOD? Maybe some info there.
This is not the same BSOD I had earlier.
http://debian.merseine.nu/bsod.png
I think this is:
http://debian.merseine.nu/bsod0.png
--
Cheers
John
-- spambait
1aaaaaaa(a)coco.merseine.nu Z1aaaaaaa(a)coco.merseine.nu
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375
You cannot reply off-list:-)
10:58 a.m.
John Summerfield wrote:
>>
> Can you capture a screenshot of the BSOD? Maybe some info there.
>
This is not the same BSOD I had earlier.
http://debian.merseine.nu/bsod.png
This one is due to the -no-acpi switch being passed to qemu, while the
guest really wants ACPI. Don't pass -no-acpi, but do look at
http://kvm.qumranet.com/kvmwiki/Windows_ACPI_Workaround.
I think this is:
http://debian.merseine.nu/bsod0.png
This is a corrupted disk, or perhaps the IDE driver is not installed
correctly.
--
error compiling committee.c: too many arguments to function
7:21 a.m.
Avi Kivity wrote:
John Summerfield wrote:
>>>
>> Can you capture a screenshot of the BSOD? Maybe some info there.
>>
>
> This is not the same BSOD I had earlier.
> http://debian.merseine.nu/bsod.png
This one is due to the -no-acpi switch being passed to qemu, while the
guest really wants ACPI. Don't pass -no-acpi, but do look at
http://kvm.qumranet.com/kvmwiki/Windows_ACPI_Workaround.
the f7 key works for that.
>
> I think this is:
> http://debian.merseine.nu/bsod0.png
>
This is a corrupted disk, or perhaps the IDE driver is not installed
correctly.
the disk is fine, I can boot it natively. The hardware emulation in qemu
is very old (circa Pentium Pro CPUs). It's possible (probable, do I
remember LBA supports 120 Gbytes? wasn't there an earlier max of 32
Gbytes?) the disk (320 Gbytes) is larger than the emulation supports.
It's also possible that Windows is picking up garbage device geometry.
--
Cheers
John
-- spambait
1aaaaaaa(a)coco.merseine.nu Z1aaaaaaa(a)coco.merseine.nu
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375
You cannot reply off-list:-)
9:39 a.m.
John Summerfield wrote:
>
>>
>> I think this is:
>> http://debian.merseine.nu/bsod0.png
>>
>
> This is a corrupted disk, or perhaps the IDE driver is not installed
> correctly.
the disk is fine, I can boot it natively. The hardware emulation in
qemu is very old (circa Pentium Pro CPUs). It's possible (probable, do
I remember LBA supports 120 Gbytes? wasn't there an earlier max of 32
Gbytes?) the disk (320 Gbytes) is larger than the emulation supports.
It's also possible that Windows is picking up garbage device geometry.
Ah, I haven't tried very large disks. I'll check it out.
--
error compiling committee.c: too many arguments to function
9:53 a.m.
On Thu, Dec 27, 2007 at 05:39:22PM +0200, Avi Kivity wrote:
John Summerfield wrote:
>
>>
>>>
>>>I think this is:
>>>http://debian.merseine.nu/bsod0.png
>>>
>>
>>This is a corrupted disk, or perhaps the IDE driver is not installed
>>correctly.
>
>the disk is fine, I can boot it natively. The hardware emulation in
>qemu is very old (circa Pentium Pro CPUs). It's possible (probable, do
>I remember LBA supports 120 Gbytes? wasn't there an earlier max of 32
>Gbytes?) the disk (320 Gbytes) is larger than the emulation supports.
>
>It's also possible that Windows is picking up garbage device geometry.
>
Ah, I haven't tried very large disks. I'll check it out.
Funnily enough, a change was just commited to the Xen's copy of the QEMU
ROMBIOS to support disks > 128 GB which I suspect would also apply to
upstream QEMU / KVM's rombios
changeset: 16669:7fbc521b07a9
tag: tip
user: Keir Fraser <keir.fraser(a)citrix.com>
date: Thu Dec 27 13:00:40 2007 +0000
files: tools/firmware/rombios/rombios.c
description:
x86, hvm, rombios: INT13 LBA48 support for disks bigger than 128GB.
The new limit should be 2TB.
Signed-off-by: Samuel Thibault <samuel.thibault(a)citrix.com>
http://xenbits.xensource.com/staging/xen-unstable.hg?rev/7fbc521b07a9
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 -=|
2:50 p.m.
Daniel P. Berrange wrote:
On Thu, Dec 27, 2007 at 05:39:22PM +0200, Avi Kivity wrote:
> John Summerfield wrote:
>>>> I think this is:
>>>> http://debian.merseine.nu/bsod0.png
>>>>
>>> This is a corrupted disk, or perhaps the IDE driver is not installed
>>> correctly.
>> the disk is fine, I can boot it natively. The hardware emulation in
>> qemu is very old (circa Pentium Pro CPUs). It's possible (probable, do
>> I remember LBA supports 120 Gbytes? wasn't there an earlier max of 32
>> Gbytes?) the disk (320 Gbytes) is larger than the emulation supports.
>>
>> It's also possible that Windows is picking up garbage device geometry.
>>
> Ah, I haven't tried very large disks. I'll check it out.
Funnily enough, a change was just commited to the Xen's copy of the QEMU
ROMBIOS to support disks > 128 GB which I suspect would also apply to
upstream QEMU / KVM's rombios
changeset: 16669:7fbc521b07a9
tag: tip
user: Keir Fraser <keir.fraser(a)citrix.com>
date: Thu Dec 27 13:00:40 2007 +0000
files: tools/firmware/rombios/rombios.c
description:
x86, hvm, rombios: INT13 LBA48 support for disks bigger than 128GB.
The new limit should be 2TB.
Signed-off-by: Samuel Thibault <samuel.thibault(a)citrix.com>
http://xenbits.xensource.com/staging/xen-unstable.hg?rev/7fbc521b07a9
Regards,
Dan.
Thanks, Dan, that's encouraging. When is (and is) this likely to filter
through to those of us who prefer to not patch their own software?
--
Cheers
John
-- spambait
1aaaaaaa(a)coco.merseine.nu Z1aaaaaaa(a)coco.merseine.nu
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375
You cannot reply off-list:-)
2:53 p.m.
On Fri, Dec 28, 2007 at 05:50:40AM +0900, John Summerfield wrote:
Daniel P. Berrange wrote:
>On Thu, Dec 27, 2007 at 05:39:22PM +0200, Avi Kivity wrote:
>>John Summerfield wrote:
>>>>>I think this is:
>>>>>http://debian.merseine.nu/bsod0.png
>>>>>
>>>>This is a corrupted disk, or perhaps the IDE driver is not installed
>>>>correctly.
>>>the disk is fine, I can boot it natively. The hardware emulation in
>>>qemu is very old (circa Pentium Pro CPUs). It's possible (probable, do
>>>I remember LBA supports 120 Gbytes? wasn't there an earlier max of 32
>>>Gbytes?) the disk (320 Gbytes) is larger than the emulation supports.
>>>
>>>It's also possible that Windows is picking up garbage device geometry.
>>>
>>Ah, I haven't tried very large disks. I'll check it out.
>
>Funnily enough, a change was just commited to the Xen's copy of the QEMU
>ROMBIOS to support disks > 128 GB which I suspect would also apply to
>upstream QEMU / KVM's rombios
>
> changeset: 16669:7fbc521b07a9
> tag: tip
> user: Keir Fraser <keir.fraser(a)citrix.com>
> date: Thu Dec 27 13:00:40 2007 +0000
> files: tools/firmware/rombios/rombios.c
> description:
> x86, hvm, rombios: INT13 LBA48 support for disks bigger than 128GB.
> The new limit should be 2TB.
>
> Signed-off-by: Samuel Thibault <samuel.thibault(a)citrix.com>
>
>
>http://xenbits.xensource.com/staging/xen-unstable.hg?rev/7fbc521b07a9
>
>Regards,
>Dan.
Thanks, Dan, that's encouraging. When is (and is) this likely to filter
through to those of us who prefer to not patch their own software?
When Avi releases -59 we'll include that in rawhide. If you need the
LBA-48 fix applied to other releases please file a BZ against the
appropriate Fedora release and add a link to this thread in the BZ
report.
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 -=|
5:20 p.m.
Daniel P. Berrange wrote:
On Fri, Dec 28, 2007 at 05:50:40AM +0900, John Summerfield wrote:
> Daniel P. Berrange wrote:
>> On Thu, Dec 27, 2007 at 05:39:22PM +0200, Avi Kivity wrote:
>>> John Summerfield wrote:
>>>>>> I think this is:
>>>>>> http://debian.merseine.nu/bsod0.png
>>>>>>
>>>>> This is a corrupted disk, or perhaps the IDE driver is not installed
>>>>> correctly.
>>>> the disk is fine, I can boot it natively. The hardware emulation in
>>>> qemu is very old (circa Pentium Pro CPUs). It's possible (probable,
do
>>>> I remember LBA supports 120 Gbytes? wasn't there an earlier max of 32
>>>> Gbytes?) the disk (320 Gbytes) is larger than the emulation supports.
>>>>
>>>> It's also possible that Windows is picking up garbage device
geometry.
>>>>
>>> Ah, I haven't tried very large disks. I'll check it out.
>> Funnily enough, a change was just commited to the Xen's copy of the QEMU
>> ROMBIOS to support disks > 128 GB which I suspect would also apply to
>> upstream QEMU / KVM's rombios
>>
>> changeset: 16669:7fbc521b07a9
>> tag: tip
>> user: Keir Fraser <keir.fraser(a)citrix.com>
>> date: Thu Dec 27 13:00:40 2007 +0000
>> files: tools/firmware/rombios/rombios.c
>> description:
>> x86, hvm, rombios: INT13 LBA48 support for disks bigger than 128GB.
>> The new limit should be 2TB.
>>
>> Signed-off-by: Samuel Thibault <samuel.thibault(a)citrix.com>
>>
>>
>> http://xenbits.xensource.com/staging/xen-unstable.hg?rev/7fbc521b07a9
>>
>> Regards,
>> Dan.
> Thanks, Dan, that's encouraging. When is (and is) this likely to filter
> through to those of us who prefer to not patch their own software?
When Avi releases -59 we'll include that in rawhide. If you need the
LBA-48 fix applied to other releases please file a BZ against the
appropriate Fedora release and add a link to this thread in the BZ
report.
Where?
I couldn't find a matching one at RH, though I saw some that could have
been created by this problem, if the disks had been big enough, and was
contemplating creating one until this happened:
Service Temporarily Unavailable
The server is temporarily unable to service your request due to
maintenance downtime or capacity problems. Please try again later
Regards,
Dan.
--
Cheers
John
-- spambait
1aaaaaaa(a)coco.merseine.nu Z1aaaaaaa(a)coco.merseine.nu
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375
You cannot reply off-list:-)
9:59 a.m.
Avi Kivity wrote:
>
> the disk is fine, I can boot it natively. The hardware emulation in
> qemu is very old (circa Pentium Pro CPUs). It's possible (probable,
> do I remember LBA supports 120 Gbytes? wasn't there an earlier max of
> 32 Gbytes?) the disk (320 Gbytes) is larger than the emulation supports.
>
Ah, I haven't tried very large disks. I'll check it out.
You were right, it was missing LBA48 support. I added it for kvm-59 (by
coincidence, an LBA48 patch was posted for Xen four hours ago; so I used
that).
I now have Windows happily installing into a 300GB virtual disk.
--
error compiling committee.c: too many arguments to function
8:55 a.m.
On Tue, 18 Dec 2007 17:58:24 +0900
John Summerfield <debian(a)herakles.homelinux.org> wrote:
John Summerfield wrote:
> Richard W.M. Jones wrote:
>> John Summerfield wrote:
>
> The good news is that the system now boots grub sans menu. I typed
> in the necessary information to boot windows:
> chainloader (hd0,0)+1
> boot
>
> and windows boots all the way to a BSOD and the stuff on the screen
> recommends "chkdsk /f."
>
> Hmm. Type to find a rescue system.
fwiw same happens when I boot directly from the partition. Ideas
welcome.
Windows is picky about hard disk drivers. The initial installation
seems to install some drivers that get loaded early for your disk
hardware. QEMU emulates a different type of hardware than what you
probably have and so it can't find the root disk once the kernel has
booted. If you change the main disk controller on a real machine, you'll
probably have the same problem.
I'm not aware of a way to fix that, unfortunately. I've generally had to
install my windows images from scratch.
--
Jeff Layton <jlayton(a)redhat.com>
9:07 a.m.
If you have this system set up how you want and it required a lot of
configuration, one option might be to do a full backup (including system
state) of the configured system, then install windows on an hvm and do a
full restore, I have read that this is a good p2v (and vice versa) method
for Windows, as the system state backup apparently doesn't replace drivers.
I have never tried it, so I can't tell you the caveats, but depending on
your situation, it might be worth a shot (I have read similar articles about
moving from one piece of hardware to another). Also, in Windows, all
drivers are loaded the same way, so if the right driver can be installed
from the initial hardware configuration, Windows will boot in the new
configuration. I have done this many times to change HD controllers in the
same system for performance reasons (going to a separate IDE controller was
a big performance boost on old IDE computers). I never had much luck doing
that, though, unless I could put the hardware in to install the drivers for,
that may mean that it won't automatically use a driver just because it is
there, it has to be associated with the specific piece of hardware.
Nonetheless, depending on how much time you want to spend, you might have
some options.
Dustin
-----Original Message-----
From: fedora-xen-bounces(a)redhat.com [mailto:fedora-xen-bounces@redhat.com]
On Behalf Of Jeff Layton
Sent: Wednesday, December 19, 2007 09:55
To: John Summerfield
Cc: Fedora Xen
Subject: Re: [Fedora-xen] kvm crash on f8
On Tue, 18 Dec 2007 17:58:24 +0900
John Summerfield <debian(a)herakles.homelinux.org> wrote:
John Summerfield wrote:
> Richard W.M. Jones wrote:
>> John Summerfield wrote:
>
> The good news is that the system now boots grub sans menu. I typed
> in the necessary information to boot windows:
> chainloader (hd0,0)+1
> boot
>
> and windows boots all the way to a BSOD and the stuff on the screen
> recommends "chkdsk /f."
>
> Hmm. Type to find a rescue system.
fwiw same happens when I boot directly from the partition. Ideas
welcome.
Windows is picky about hard disk drivers. The initial installation
seems to install some drivers that get loaded early for your disk
hardware. QEMU emulates a different type of hardware than what you
probably have and so it can't find the root disk once the kernel has
booted. If you change the main disk controller on a real machine, you'll
probably have the same problem.
I'm not aware of a way to fix that, unfortunately. I've generally had to
install my windows images from scratch.
--
Jeff Layton <jlayton(a)redhat.com>
--
Fedora-xen mailing list
Fedora-xen(a)redhat.com
https://www.redhat.com/mailman/listinfo/fedora-xen
6:57 p.m.
Dustin Henning wrote:
If you have this system set up how you want and it required a lot
of
configuration, one option might be to do a full backup (including system
state) of the configured system, then install windows on an hvm and do a
full restore, I have read that this is a good p2v (and vice versa) method
for Windows, as the system state backup apparently doesn't replace drivers.
I have never tried it, so I can't tell you the caveats, but depending on
your situation, it might be worth a shot (I have read similar articles about
moving from one piece of hardware to another). Also, in Windows, all
drivers are loaded the same way, so if the right driver can be installed
from the initial hardware configuration, Windows will boot in the new
configuration. I have done this many times to change HD controllers in the
same system for performance reasons (going to a separate IDE controller was
a big performance boost on old IDE computers). I never had much luck doing
that, though, unless I could put the hardware in to install the drivers for,
that may mean that it won't automatically use a driver just because it is
there, it has to be associated with the specific piece of hardware.
Nonetheless, depending on how much time you want to spend, you might have
some options.
Dustin
<snip>
and now I think it's Jeff.
Windows is picky about hard disk drivers. The initial installation
seems to install some drivers that get loaded early for your disk
hardware. QEMU emulates a different type of hardware than what you
probably have and so it can't find the root disk once the kernel has
booted. If you change the main disk controller on a real machine, you'll
probably have the same problem.
I'm not aware of a way to fix that, unfortunately. I've generally had to
install my windows images from scratch.
This is a trial of what I really want to do, which includes running
Windows 2003 Server and a domain of Windows XP Professional clients (not
many - likely 2-3 of them) for testing and education, and ideally a
Windows XP system which is to be moved to different hardware anyway, but
preferably without being reinstalled.
Installing drivers is okay, using xen is fine, I tried kvm because it
doesn't require a special kernel.
I'm a bit dismayed to see mentions of FX chipset and piix3 in the
qemu-kvm doc (I remember piix4, but piix3?). Surely that won't work for
Vista.
I did try an install of my Windows on the HP dc7700p using the ISO in
its recovery partition, but I quit as soon as it asked me for the other
CD as evidence I'm licenced[1]. I suspect this version is locked to my
BIOS and so doesn't require activation, so it's probably not going to
work at all, but enquiring minds need to know.
The backup/restore procedure might work, it's supposed to preserve user
data, but if it's supposed to clean out malware then it should not.
I might let it sit though, and see whether someone has another idea:-)
Merry Xmas all.
[1] Producing the CD isn't a problem, sticking it in the imaginary CD
drive was, and while I think the needed ISO is right there too, I
especially don't know how to shuffle virtual CDs.
--
Cheers
John
-- spambait
1aaaaaaa(a)coco.merseine.nu Z1aaaaaaa(a)coco.merseine.nu
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375
You cannot reply off-list:-)
11:38 a.m.
On Thu, 20 Dec 2007 09:57:24 +0900
John Summerfield <debian(a)herakles.homelinux.org> wrote:
[1] Producing the CD isn't a problem, sticking it in the imaginary CD
drive was, and while I think the needed ISO is right there too, I
especially don't know how to shuffle virtual CDs.
Generally, what I've had to do is run qemu-kvm from the command line
with "-monitor stdio". When you need to change cd's, in the qemu
monitor console you do:
eject cdrom
change cdrom /path/to/iso/image.iso
It would be nice if there were some way to do this from virsh or the
virt-manager GUI, but I don't think there is as of yet. Maybe open a BZ
for it if you're feeling daring.
When I do this, I usually disable networking since libvirt isn't
setting it up anyway (-net none). Once the image is installed, you can
boot it as usual via virt-manager or virsh with networking enabled.
--
Jeff Layton <jlayton(a)redhat.com>
4:46 p.m.
Jeff Layton wrote:
On Thu, 20 Dec 2007 09:57:24 +0900
John Summerfield <debian(a)herakles.homelinux.org> wrote:
> [1] Producing the CD isn't a problem, sticking it in the imaginary CD
> drive was, and while I think the needed ISO is right there too, I
> especially don't know how to shuffle virtual CDs.
>
Generally, what I've had to do is run qemu-kvm from the command line
with "-monitor stdio". When you need to change cd's, in the qemu
monitor console you do:
Most of my vms are xen; I decided to give kvm a fly when I decided I
really don't like that I can't get framebuffer working on the xenified
kernel.
If kvm works, that's probably what I'll use where I have hardware
assist, I like that I don't have to choose it at boot tine.
--
Cheers
John
-- spambait
1aaaaaaa(a)coco.merseine.nu Z1aaaaaaa(a)coco.merseine.nu
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375
You cannot reply off-list:-)
5961
days inactive
5973
days old
18 comments
6 participants
participants (6)
-
Avi Kivity -
Daniel P. Berrange -
Dustin Henning -
Jeff Layton -
John Summerfield -
Richard W.M. Jones