[ovirt-users] Unable to start VM after upgrade vom 4.1.9 to 4.2.1 - NPE

Oliver Riesener Oliver.Riesener at hs-bremen.de
Wed Mar 7 16:51:18 UTC 2018


On 07.03.2018 17:22, Arik Hadas wrote:
>
>
> On Wed, Mar 7, 2018 at 6:15 PM, Jan Siml <jsiml at plusline.net 
> <mailto:jsiml at plusline.net>> wrote:
>
>     Hello,
>
>             Enable network and disks on your VM than do:
>             Run -> ONCE Ok Ignore errors. Ok
>             Run
>             Cheeers
>
>
>         WTF! That worked.
>
>         Did you know, why this works and what happens in the
>         background? Is there a Bugzilla bug ID for this issue?
>
I figured it out, by they attempt to change the VM CPU family from old 
VM's, as a last try to get it work again.
After i had a live upgrade to 4.2 with running VM's and shutdown them 
down they all dead with disabled network and disks.
No delight to delete them all and recreate them, the other way.

Have a nice day.

>     BTW, here is the list of devices before:
>
>     engine=# select type, device, address, is_managed, is_plugged,
>     alias from vm_device where vm_id in (select vm_guid from vm_static
>     where vm_name='prod-hub-201');
>         type    |    device     |  address                    |
>     is_managed | is_plugged |    alias
>     ------------+---------------+--------------------------------------------------------------+------------+------------+----------------
>      video      | qxl           |                    | t       | t   
>           |
>      controller | virtio-scsi   |                    | t       | t   
>           |
>      balloon    | memballoon    |                    | t       | f   
>           | balloon0
>      graphics   | spice         |                    | t       | t   
>           |
>      controller | virtio-serial | {slot=0x06, bus=0x00, domain=0x0000,
>     type=pci, function=0x0} | t          | t       | virtio-serial0
>      disk       | disk          | {slot=0x07, bus=0x00, domain=0x0000,
>     type=pci, function=0x0} | f          | t       | virtio-disk0
>      memballoon | memballoon    | {slot=0x08, bus=0x00, domain=0x0000,
>     type=pci, function=0x0} | f          | t       | balloon0
>      interface  | bridge        | {slot=0x03, bus=0x00, domain=0x0000,
>     type=pci, function=0x0} | f          | t       | net0
>      interface  | bridge        | {slot=0x09, bus=0x00, domain=0x0000,
>     type=pci, function=0x0} | f          | t       | net1
>      controller | scsi          | {slot=0x05, bus=0x00, domain=0x0000,
>     type=pci, function=0x0} | f          | t       | scsi0
>      controller | ide           | {slot=0x01, bus=0x00, domain=0x0000,
>     type=pci, function=0x1} | f          | t         | ide
>      controller | usb           | {slot=0x01, bus=0x00, domain=0x0000,
>     type=pci, function=0x2} | t          | t         | usb
>      channel    | unix          | {bus=0, controller=0,
>     type=virtio-serial, port=1}            | f          | t         |
>     channel0
>      channel    | unix          | {bus=0, controller=0,
>     type=virtio-serial, port=2}            | f          | t         |
>     channel1
>      channel    | spicevmc      | {bus=0, controller=0,
>     type=virtio-serial, port=3}            | f          | t         |
>     channel2
>      interface  | bridge        |                    | t       | t   
>           | net1
>      interface  | bridge        |                    | t       | t   
>           | net0
>      disk       | cdrom         |                    | t       | f   
>           | ide0-1-0
>      disk       | cdrom         | {bus=1, controller=0, type=drive,
>     target=0, unit=0}          | f          | t         | ide0-1-0
>      disk       | disk          |                    | t       | t   
>           | virtio-disk0
>     (20 rows)
>
>     and afterwards:
>
>     engine=# select type, device, address, is_managed, is_plugged,
>     alias from vm_device where vm_id in (select vm_guid from vm_static
>     where vm_name='prod-hub-201');
>         type    |    device     |  address                    |
>     is_managed | is_plugged |    alias
>     ------------+---------------+--------------------------------------------------------------+------------+------------+----------------
>      channel    | spicevmc      | {type=virtio-serial, bus=0,
>     controller=0, port=3}            | f          | t | channel2
>      channel    | unix          | {type=virtio-serial, bus=0,
>     controller=0, port=1}            | f          | t | channel0
>      interface  | bridge        | {type=pci, slot=0x04, bus=0x00,
>     domain=0x0000, function=0x0} | t          | t       | net1
>      controller | usb           | {type=pci, slot=0x01, bus=0x00,
>     domain=0x0000, function=0x2} | t          | t       | usb
>      controller | virtio-serial | {type=pci, slot=0x06, bus=0x00,
>     domain=0x0000, function=0x0} | t          | t       | virtio-serial0
>      interface  | bridge        | {type=pci, slot=0x03, bus=0x00,
>     domain=0x0000, function=0x0} | t          | t       | net0
>      controller | virtio-scsi   | {type=pci, slot=0x05, bus=0x00,
>     domain=0x0000, function=0x0} | t          | t       | scsi0
>      video      | qxl           | {type=pci, slot=0x02, bus=0x00,
>     domain=0x0000, function=0x0} | t          | t       | video0
>      channel    | unix          | {type=virtio-serial, bus=0,
>     controller=0, port=2}            | f          | t | channel1
>      balloon    | memballoon    |                    | t     | t     
>         | balloon0
>      graphics   | spice         |                    | t     | t     
>         |
>      disk       | cdrom         |                    | t       | f   
>           | ide0-1-0
>      disk       | disk          | {type=pci, slot=0x07, bus=0x00,
>     domain=0x0000, function=0x0} | t          | t       | virtio-disk0
>     (13 rows)
>
>
> Thanks.
> The problem was that unmanaged interfaces and disks were created (and 
> thus, you previously had 4 interfaces devices, 2 disk devices and 2 CD 
> devices).
> That is most probably a result of a bug we had when migrating a VM 
> that was started in cluster < 4.2 to 4.2 host.
> The fix for this bug will be available in 4.2.2.
> You could, alternatively, remove the unmanaged (disk and interface) 
> devices and plug the managed ones.
>
>     Regards
>
>     Jan
>
>

-- 
Mit freundlichem Gruß


Oliver Riesener

--
Hochschule Bremen
Elektrotechnik und Informatik
Oliver Riesener
Neustadtswall 30
D-28199 Bremen

Tel: 0421 5905-2405, Fax: -2400
e-mail:oliver.riesener at hs-bremen.de

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20180307/2af5101f/attachment.html>


More information about the Users mailing list