[Users] Problem on boot after deleting last snaphot

[Gianluca Cecchi]

Hello,
snapshot testing in 3.2.1 on f18 host with
http://gerrit.ovirt.org/#/c/13172/
(enable template from snapshot)

- create one vm named prova_snap from centos 6 template
and choose preallocated storage

- power on

- create test_snap1 with vm powered on
Snapshot test_snap1 creation for VM prova_snap has been completed.
I have now:

Current OK Active VM
2013-Apr-16, 14:09 OK test_snap1


- create some files

- shutdown VM
VM shutdown initiated by admin at internal on VM prova_snap (Host: f18ovn03).
VM prova_snap is down. Exit message: User shut down

- snapshots --> select "test_snap1" --> preview
Snapshot-Preview for VM prova_snap was initiated by admin at internal.
Snapshot-Preview for VM prova_snap has been completed.

And I have under "Description" field of snapshots view
test_snap1 (Preview Mode) (in coloured bold)
2013-Apr-16, 14:09 In Preview test_snap1 (Preview Mode)

- Power on VM
--> ok the files created after the snapshot are not there.

- Shutdown VM and undo the operation to come back (forward actually)
to the active state
Restoring VM prova_snap from snapshot started by user admin at internal.
VM prova_snap restoring from Snapshot has been completed.

- Power On VM
Ok, all files are there again.

- create test_snap2 with vm powered on
Snapshot test_snap2 creation for VM prova_snap was initiated by admin at internal.
Snapshot test_snap2 creation for VM prova_snap has been completed.

Now I have
Current OK Active VM
2013-Apr-16, 16:21 OK test_snap2
2013-Apr-16, 14:09 OK test_snap1

- create some files
/testfile_aftersnap2

- shutdown VM

- snapshots --> select again the first snapshot named "test_snap1" --> preview
Snapshot-Preview for VM prova_snap was initiated by admin at internal.
Snapshot-Preview for VM prova_snap has been completed.

Now I have:
2013-Apr-16, 16:25 OK Active VM before the preview
2013-Apr-16, 16:21 OK test_snap2
2013-Apr-16, 14:09 In Preview test_snap1 (Preview Mode)

- Power on VM
--> ok the files created after the snapshots are not there.

- Shutdown VM and commit so that subsequent snapshots (in my case
test_snap2) should be erased
Restoring VM prova_snap from snapshot started by user admin at internal.
VM prova_snap restoring from Snapshot has been completed.

Now I correctly have only this:
Current OK Active VM
2013-Apr-16, 14:09 OK test_snap1

- Power on VM
And now I should be in situation as right after test_snap1 creation if
I power on te VM
--> OK


Now I would like to delete the remaining snapshot (test_snap1) and
begin a new story...

So:

- Power off the VM

- Delete snapshot
Are you sure you want to delete snapshot from Tue Apr 16 14:09:10
GMT+200 2013 with description 'test_snap1'?
yes

during the process
Total DISK READ:   46660.26 K/s | Total DISK WRITE:   46468.59 K/s
  PID  PRIO  USER    DISK READ>  DISK WRITE  SWAPIN      IO    COMMAND
20610 idle vdsm     46036.93 K/s 46457.74 K/s  0.00 % 30.67 % qemu-img
convert -t n~cb-2eb88d4ede9a_MERGE
 2457 be/4 sanlock   336.65 K/s    0.00 K/s  0.00 %  0.03 % sanlock
daemon -U sanlock -G sanlock
 4173 be/3 vdsm        0.33 K/s    0.00 K/s  0.00 %  0.00 % python
/usr/share/vdsm~eFileHandler.pyc 43 40
 2830 be/4 root        0.00 K/s   11.84 K/s  0.00 %  0.00 % libvirtd --listen
  875 be/3 root        0.00 K/s    0.00 K/s  0.00 %  0.52 % [jbd2/dm-12-8]
 3141 be/3 vdsm        0.00 K/s    1.32 K/s  0.00 %  0.00 % python
/usr/share/vdsm/vdsm

ps:

vdsm     20610  3141  7 16:43 ?        00:00:01 /usr/bin/qemu-img
convert -t none -f qcow2
/rhev/data-center/5849b030-626e-47cb-ad90-3ce782d831b3/3fb66ba1-cfcb-4341-8960-46f0e8cf6e83/images/78c29803-45d5-46fb-bdd6-39647c067c28/b1e4782f-ff51-4a25-94cb-2eb88d4ede9a
-O raw /rhev/data-center/5849b030-626e-47cb-ad90-3ce782d831b3/3fb66ba1-cfcb-4341-8960-46f0e8cf6e83/images/78c29803-45d5-46fb-bdd6-39647c067c28/b1e4782f-ff51-4a25-94cb-2eb88d4ede9a_MERGE

Snapshot test_snap1 deletion on VM prova_snap was initiated by admin at internal.
(in tasks:
executing
    --> Merging snapshots of disk prova_snap_Disk1
)
Snapshot test_snap1 deletion on VM prova_snap has been completed.
(elapsed about 4 minutes)

Now in snapshots pane:
Current OK Active VM

But the VM doesn't boot, so it seems that the final operation
corrupted the MBR perhaps....

see this screenshot where there is console window and disks pane:
https://docs.google.com/file/d/0BwoPbcrMv8mvRDlLZEw1dmlvM0U/edit?usp=sharing

Any hint? Anyone can replicate and reproduce?

Thanks
Gianluca