Gonzalo Pérez de Olaguer Córdoba a écrit :
> Package: qemu
> Version: 0.10.6-1
> Followup-For: Bug #411814
>
>
> I'm asking to reopen this bug because the symptoms are quite the same.
>
> I'm using the version of kqemu and dependencies from squeeze, although
> the base system is lenny.
>
> Trying to install debian into a qemu image using the lenny netinst disk
> image some parts of the installation (disk formatting and
> package installation) go veru slow, and eventually dumps core (segmentation
> fault). This only happens when using qcow images.
>
> When using images in raw format everything goes fast.
>
> I suspect this is a problem related to disk write access in qcow format.
>
> For your information, I'm creating images this way:
>
> qemu-img -f qcow lenny-base.qcow 10G
>
> and running qemu this way:
>
> sudo qemu -m 256 -net nic,vlan=1,macaddr=52:54:00:12:34:51 -net tap,vlan=1
> -cdrom debian-502a-i386-netinst.iso -boot d lenny-base.qcow
>
I am almost sure it's a different problem. I guess your problems are due
to cache policy as shown in the NEWS.Debian file (see below).
You should try using '-drive file=lenny-base.qcow,cache=writeback'
instead of simply 'lenny-base.qcow'.
Aurelien
qemu (0.10.3-2) unstable; urgency=low
Starting with QEMU 0.10.0, it is possible to control how the host
cache is used to access block data, using the cache= suboption of the
-drive option. The following suboptions are available:
* none: The host page cache is entirely avoided.
* writeback (default in QEMU 0.9.x): Writeback caching reports data
writes as completed as soon as the data is present in the host page
cache. This is safe as long as you trust your host. If your host
crashes or loses power, then the guest may experience data
corruption.
* writethrough (default in QEMU 0.10.x): The host page cache is used
to read and write data but write notification is sent to the guest
only when the data has been reported as written by the storage
subsystem.
Note that depending on your configuration (filesystem, encryption,
kernel version, etc.), disk accesses can be very slow with the default
cache policy (writethrough). You can use the writeback cache policy
instead, but the data integrity is not assured anymore.
See qemu(1) for more details.
-- Aurelien Jarno <[email protected]> Sun, 03 May 2009 23:22:29 +0200
--
Aurelien Jarno GPG: 1024D/F1BCDB73
[email protected] http://www.aurel32.net
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
