Appvm freezes whole system on startup?!

[Stumpy]

I shutdown an appvm then tried to start it back up, but now, when i try
to start it up, my monitor goes white (with a very thin outline of the
appvm that i started, and then nothing. I cant kill it, change
workspaces, access menus, nada.

I have tried to open a few things on that appvm but no matter what it
seems to crash whether its FF or xterm etc, all freeze my sys at which
point i have to restart.

How can i recover from this? or at least salvage the data in this appvm?

I am running a fully updated ver of qubes, the appvm is deb10.

[Stumpy]

UPDATE:
Now another appvm is freezing my system, also deb10.

I thought Appvms shouldnt be able do freeze a whole qubes installation?

I havent done anything with dom0 recently other than updating it but
this is getting serious as the previous appvm was not a huge deal (major
inconvenience though) but this more recent one is a big deal as its for
work.

Suggestions really needed. I'd be happy to post whatever info/log file
needed, i just dont know which one would be useful.

[unman]

I'm assuming you havent allocated crucial device to that qube?
You can create new, and recover the data:
`ls -l /dev/mapper/qubes_dom0-vm--broken--private` to get name of block
device. (dm-XXX)
`qvm-start new --hddisk dom0:/dev/dm-XXX`

In new: mount,copy data across,umount, shutdown
(optional) delete broken qube.

[unman]

Dont use deb10 until you've secured data.

[Stumpy]

On 2020-03-06 08:59, unman wrote:
> On Fri, Mar 06, 2020 at 08:16:20AM -0500, Stumpy wrote:
>> I shutdown an appvm then tried to start it back up, but now, when i try to
>> start it up, my monitor goes white (with a very thin outline of the appvm
>> that i started, and then nothing. I cant kill it, change workspaces, access
>> menus, nada.
>>
>> I have tried to open a few things on that appvm but no matter what it seems
>> to crash whether its FF or xterm etc, all freeze my sys at which point i
>> have to restart.
>>
>> How can i recover from this? or at least salvage the data in this appvm?
>>
>> I am running a fully updated ver of qubes, the appvm is deb10.
>>
>
> I'm assuming you havent allocated crucial device to that qube?

There was a device attached when i shut it down (before the issues
started) but now its not showing as attached to anything.

The other appvm didnt have anything attached to it but its started
acting the same, that is freezing the system when i start it up.

> You can create new, and recover the data:
> `ls -l /dev/mapper/qubes_dom0-vm--broken--private` to get name of block
> device. (dm-XXX)
> `qvm-start new --hddisk dom0:/dev/dm-XXX`
>
> In new: mount,copy data across,umount, shutdown
> (optional) delete broken qube.
>

Thanks, I can try that, but since it seems to be starting in other
appvms i am now not sure if this is an appvm thing or a dom0 thing?

[Stumpy]

On 2020-03-06 08:16, Stumpy wrote:

I have checked and confirmed that all my deb10 appvms freeze my system.
While i can recover the data I really need to have working appvms as
there are some things that are harder to run on fedora than deb.

Is anyone else having a similar problem?

I even tried to install a new template but got this:
@dom0 ~]$ sudo qubes-dom0-update --enablerepo=qubes-templates-itl
qubes-template-debian-10
Using sys-whonix as UpdateVM to download updates for Dom0; this may take
some time...
No Match for argument qubes-template-debian-10
Nothing to download


Any help on how to diagnose/resolve this would really be appreciated as
i really dont know where to start.

[dono...@unseen.is]

> --
I had a similar issue re-installing the deb template and I had success using "sudo qubes-dom0-update --action=reinstall qubes-template-debian-10".

[Stumpy]

Thanks for that. Just to check though, would that remove my current
template? or keep the current and install a new one?

[Sven Semmler]

On Sat, Mar 07, 2020 at 06:15:11AM -0500, Stumpy wrote:
> On 2020-03-06 08:16, Stumpy wrote:
> > I shutdown an appvm then tried to start it back up, but now, when i try
> > to start it up, my monitor goes white (with a very thin outline of the
> > appvm that i started, and then nothing. I cant kill it, change
> > workspaces, access menus, nada.
>

> Is anyone else having a similar problem?

I had a simlilar issue months ago when I installed 'gnome' as in the
whole desktop environment. I suggest when doing so one can damage the
GUI integration with Qubes.

When installing packages or trying things out I've never done before I
always do so in an AppVM, no the TemplateVM. The advantage being, if I
mess things up all my changes to the VM are gone once I restart it.

Only after testing changes successfully in an AppVM do I then install
them in the TemplateVM (with the obvious exception of things I've
installed before or that couldn't conceivably mess things up).

/Sven

--
public key: https://www.svensemmler.org/0x8F541FB6.asc
fingerprint: D7CA F2DB 658D 89BC 08D6 A7AA DA6E 167B 8F54 1FB6

[dono...@unseen.is]

> --
I do not know as I had deleted my deb template to make space on my primary drive. You might clone your existing (to be safe if you need anything from it) and see if it does delete the original on reinstall. I do know my install squawked because it detected that there was not an existing deb template domain. However, the install ran anyway, so it may in fact delete the existing domain.

[Stumpy]

On 2020-03-06 08:16, Stumpy wrote:

Welll I finally figured out which was the offending application, anydesk
(I know, close source etc, shame on me). I added the anydesk repo and
then when i installed it, BANG, system froze - so, I'd suggest avoiding
anydesk for now at least.

I do have a final question though, how can an app effectively crash the
whole system? Ok, the, in the case "Fedora" template was borked, but not
dom0 so should a template and its appvms be able to bring the system
down like that?