Re: [Qemu-devel] QCOW(2) image corruption under QEMU 0.9.0 reproducible

[herbie hancock <hhancock@web.de>]

> -----Ursprüngliche Nachricht-----
> Von: qemu-devel@nongnu.org
> Gesendet: 16.03.07 13:11:15
> An: qemu-devel@nongnu.org
> Betreff: Re: [Qemu-devel] QCOW(2) image corruption under QEMU 0.9.0 reproducible


> herbie hancock wrote:
> > Hello, i had also a reproducible disk crash:
> > info of the last good image, size is about 3,5GB
> > 
> > I never experienced such a bad problem with qemu before, maybe it is a problem with qcow2 format ?
> 
> After the problems with qcow2 images which I reported here a few weeks
> ago, I've only been using qcow images (under QEMU 0.9.0), without such
> surprises.  So it seems qemu has some bug related to qcow2 images,
> maybe manifesting itself only after they get larger than 4GB...
> 
> 
> Best regards
>                           J Esteves
> 
> 
> _______________________________________________
> Qemu-devel mailing list
> Qemu-devel@nongnu.org
> http://lists.nongnu.org/mailman/listinfo/qemu-devel
> 

I am quite sure now that it is a problem with the qcow2 format. 
I can reprodue the bug only on a qcow2 image. I did a disk fill test with dd on a qcow2 Image and the same test on a qcow Image, 
both images had  the same content. (Image 16GB Debian 4.0 Etch, hda1 ext3 15 gb,  hda5 swap 700MB)

The qcow2 Image was totally corrupted and is not bootable any more, the qcow Image passed the tests without any errors
Host system was now winxp sp2  instead of win2k, so i dont think its a problem of the host os.

I think a big  warning for the users is necessary that they should work with qcow instead of qcow2 until the bug is fixed

Best regards
hr

Test Setup:
---------------
Made a copy of the last known good qcow2 image (size 3.3 GB) and converted it to qcow

..\QemuManager\qemu\qemu-img.exe convert debian4_0.dsk -O qcow debian4_0_qcow.dsk

Crashtest with debian4_0.dsk (qcow2 format)
-----------------------------------------------------------

Size of intact debian4_0.dsk 3.551.580.160 Bytes

file format: qcow2
virtual size: 16G (16777216000 bytes)
disk size: 3.3G
cluster_size: 4096
Snapshot list:
ID        TAG                 VM SIZE                DATE       VM CLOCK
1         test                    79M 2007-03-11 18:34:49   00:07:47.475


Start VM: I:\debian4.0\QemuManager\qemu\qemu.exe -L "I:\debian4.0\QemuManager\qemu" -m 256 -hda "I:\debian4.0\debian4\debian4_0.dsk" -cdrom "I:\debian4.0\debian4\debian-testing-i386-DVD-1.iso" -kernel-kqemu -net nic,vlan=0 -net user,vlan=0 -std-vga -localtime

fsck on mounted filesystem: fsck -nv /dev/hda1 show no errors except # of free inodes
df 19% of system used (Total: 15409468 used: 2641020)
fill the rest of the disk: dd if=/dev/zero of=/tmp/bigfile bs=16M
dd stops with disk full, ok
---> now fsck -nv /dev/hda1 shows endless output of errors    <-----    ERROR
rm bigfile
shutdown system did work

---> Restart VM: Fatal not a bootable harddisk ext3 is totally corrupt    <-------  ERROR

In this case there was no crash on the VM, as i had on WIN2K, header of the image file seems to be still intact

Info after crash:
image: debian4_0.dsk
file format: qcow2
virtual size: 16G (16777216000 bytes)
disk size: 16G         <---- 16GB (why so much on a 15GB Partition ?)
cluster_size: 4096
Snapshot list:
ID        TAG                 VM SIZE                DATE       VM CLOCK
1         test                    79M 2007-03-11 18:34:49   00:07:47.475

Size of crashed debian4_0.dsk 16.659.554.304 Bytes



Test with debian4_0_qcow.dsk:
------------------------------------------

Size of debian4_0_qcow.dsk: 3.105.832.960 Bytes

image: debian4_0_qcow.dsk
file format: qcow
virtual size: 16G (16777216000 bytes)
disk size: 2.9G
cluster_size: 4096

Start VM: I:\debian4.0\QemuManager\qemu\qemu.exe -L "I:\debian4.0\QemuManager\qemu" -m 256 -hda "I:\debian4.0\debian4\debian4_0_qcow.dsk" -cdrom "I:\debian4.0\debian4\debian-testing-i386-DVD-1.iso" -kernel-kqemu -net nic,vlan=0 -net user,vlan=0 -std-vga -localtime

fsck on mounted filesystem: fsck -nv /dev/hda1 show no errors except # of free inodes
df 19% of system used (Total: 15409468 used: 2641020)
fill the rest of the disk: dd if=/dev/zero of=/tmp/bigfile bs=16M
dd stops with disk full, ok
now fsck -nv /dev/hda1 no errors
rm bigfile
shutdown system did work

image: debian4_0_qcow.dsk
file format: qcow
virtual size: 16G (16777216000 bytes)
disk size: 15G
cluster_size: 4096

Size of debian4_0_qcow.dsk: 15.919.493.120 Bytes

Start VM: all o.k.
-----------------------------------------------------



_______________________________________________________________________
Viren-Scan für Ihren PC! Jetzt für jeden. Sofort, online und kostenlos.
Gleich testen! http://www.pc-sicherheit.web.de/freescan/?mc=022222