Thanks guys, setting the VM to headless and then removing this option seemed to do the trick. Indeed, this VM was created back in 3.6, however it had QXL set in options. Seems that despite this something was not migrated correctly.

I think I have more machines in this situation, in case the oVirt team would like to make some additional tests.


El 2018-02-15 19:39, Vineet Khandpur escribió:

Just had the same issue

@ 4.1, upgraded to 4.2(.1)

Updated cluster compatibility, then, data centre compatibility

All VMs lost their hardware (NICs (showed attached but unplugged),
disks (status changed to disabled) and console)

Our solution was simply to connect the NICs, activate the disks, then
edit the VM and set Console to headless.

Shut down the VM

Then before bringing it back up, unchecked headless in the VM

We then had to do a Run-Once which failed

Then did a normal Run.

Console was available, and all hardware came back fine.

Didn't have to delete and re-create anything (although had to perform
the above on all 70+ production hosts including our main web servers
and HA load balancers .. which wasn't fun) ...

Hope this helps someone


On 15 February 2018 at 12:28, John Taylor <> wrote:

On Thu, Feb 15, 2018 at 11:54 AM, Michal Skrivanek
<> wrote:

On 15 Feb 2018, at 15:58, John Taylor <>

Hi Nicolas,
I had the same problem and it looked like it was because of some
vms (I believe from 3.6) that were configured with console with
type of CIRRUS and protocol VNC.

3.6 had cirrus indeed. That should work. Can you somehow confirm
it was really a 3.6 VM and it stopped working in 4.2? The exact
steps are important, unfortunately.

I'm pretty sure they were VMs created in 3.6, but I can't say for
absolute certain.  Sorry.

Tracing it out it showed that the vm  libvirt xml was begin set

do you at least recall what cluster level version it was when it
stopped working? The VM definition should have been changed to VGA
when you move the VM from 3.6 cluster to a 4.0+

I upgraded from 4.1.something and I'm pretty sure at the time the
cluster level was 4.1, and those same VMs were able to get
Sorry I can't be more help now. I'll see if I have any notes that
might help me remember.

  I tried different settings but the only thing that seemed
to work was to set them to headless, then reopen config and set
to something else.


On Thu, Feb 15, 2018 at 8:48 AM,  <> wrote:

We upgraded one of our infrastructures to 4.2.0 recently and
since then some
of our machines have the "Console" button greyed-out in the
Admin UI, like
they were disabled.

I changed their compatibility to 4.2 but with no luck, as
they're still

Is there a way to know why is that, and how to solve it?

I'm attaching a screenshot.

Users mailing list [1]

Users mailing list [1]

Users mailing list [1]


Users mailing list
Users mailing list

Reply via email to