From: "Gianluca Cecchi" <gianluca.cecchi(a)gmail.com>
To: "Alissa Bonas" <abonas(a)redhat.com>
Cc: "users" <users(a)ovirt.org>, "Liron Aravot"
Sent: Monday, March 4, 2013 4:33:40 PM
Subject: Re: [Users] clone vm from snapshot problem in 3.2
On Mon, Mar 4, 2013 at 1:19 PM, Alissa Bonas wrote:
>> > 1)
>> > OK, if I clone a powered on VM from a snapshot with two disks
>> > and
>> > here
>> > you can find clone in action with webadmin gui and iotop on node
>> > that
>> > shows they are cloning in parallel.. well!
>> > The VM is a slackware 14 32bit with virtio disk that I obtained
>> > from
>> > a
>> > virt-v2v from CentOS 6.3+Qemu/KVM
>> > The problem is that the cloned VM recognizes the disks in
>> > reversed
>> > order
>> > See these images where sl1432 is master slcone is the clone
>> > disk layout in details pane seems equal with boot disk the one
>> > that
>> > appears as the second, but the master boots ok, the slave no.
>> > Disks are swapped
>> > Master VM disk details:
>> > Clone VM disks details:
>> > Page with the two consoles where you can see that vda of master
>> > becomes vdb of clone and vice-versa:
>> > Can I swap again in some way? In VMware for example you can see
>> > and
>> > edit SCSI IDs of disks...
> Can you please provide the engine logs where the boot
> success/failure of master and clone can be seen?
here it is:
Starting point 15:16 both powered off.
First message with 15:17 is boot of sl1432b that is the master and
First message with 15:21 is boot of slclone that is the clone.
The slackware OS uses lilo as boot loader and on master it is
configured this way at the moment:
root@sl1432b:~# grep -v ^# /etc/lilo.conf
boot = /dev/vda
bitmap = /boot/slack.bmp
bmp-colors = 255,0,255,0,255,0
bmp-table = 60,6,1,16
bmp-timer = 65,27,0,255
timeout = 50
vga = normal
disk=/dev/vda bios=0x80 max-partitions=7
image = /boot/vmlinuz
root = /dev/vda2
initrd = /boot/initrd.gz
label = Linux
The "disk=..." entry was added to be able to support boot from device
of type vda in slackware, that doesn't support it ootb..
Thanks for additional information.
The engine log actually shows that slclone started ok:
"VM slclone c6c56d41-d70d-4b9b-a1cb-8b0c097b89a0 moved from PoweringUp -->
Can you explain what problem are you experiencing in that VM?
Also, could provide the vdsm log from the same timeframe?