On Mon, Apr 28, 2008 at 4:15 PM, Amos Shapira <[EMAIL PROTECTED]> wrote:
>  Any more ideas?

Found a solution.

>From http://www.centos.org/modules/newbb/viewtopic.php?topic_id=11685

I noticed that I also have such errors in the boot sequence:

XENBUS: Device with no driver: device/vbd/51712
XENBUS: Device with no driver: device/vbd/51744
XENBUS: Device with no driver: device/vif/0

So I followed the instructions there (only I did this in chroot on the
xen host, so "uname -r" had to be replaced with the explicit kernel
version). It didn't work for the later kernel (2.6.18-53.1.14.el5) but
with the older one (2.6.18-53) it allowed the guest to boot. The
actualy mkinitrd command which did the trick was:

mkinitrd --omit-scsi-modules --with=xennet --with=xenblk
--preload=xenblk /boot/initrd-2.6.18-53.el5xen.img 2.6.18-53.el5xen
(after moving aside the original initrd file).

I forced a re-install of the newer kernel (just erased it and did "yum
update") and I still get this XENBUS error but it boots fine. Go
figure.

BTW, to be able to do that from the xen host, I made a couple of scripts:

setup:
#!/bin/sh -ve
kpartx -a /dev/xen02/reece-test
vgchange -ay reecetest
mount /dev/reecetest/root /mnt/tmp
mount /dev/mapper/reece-test1 /mnt/tmp/boot/
mount -o bind /dev /mnt/tmp/dev
mount -o bind /proc /mnt/tmp/proc
mount -o bind /sys /mnt/tmp/sys

and teardown:
#!/bin/sh -ex

umount /mnt/tmp/boot/
umount /mnt/tmp/dev
umount /mnt/tmp/proc
umount /mnt/tmp/sys
umount /mnt/tmp
vgchange -an reecetest
kpartx -d /dev/xen02/reece-test

Which I plan to keep around as they are useful to do all sorts of
stuff with xen image LVM's. The LV itself (/dev/xen02/reece-test)
contains a partition table with one boot partition and another
partition with one large LV on it.

Cheers,

--Amos

=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]

Reply via email to