Shutdown Proccess

[Eva Star]

https://github.com/QubesOS/qubes-issues/issues/1581

What do you think about adding "priority" to each vm to solve shutdown process issue?

With highest (>=) priority number VM shutdown first and with the same number maybe synchronously.

Example:

sys-net -- shutdown priority 10

sys-firewall -- 20

sys-whonix -- 30

Some-ProxyVm with sys-whonix netvm -- auto priority 31 or manual selected 40 (Maybe user can choose priority. New vm must created with +1 shutdown priority by default or manually selected.)

When shutdown process start it must shutdown VMs with highest priority first, then continue shutdown with lower priority number VMs. 

At the same time Qubes must show fullscreen list of all VMs and status of them with buttons to kill them.

The idea to shutdown with dependencies. Like users do at QM when they wants to stop all vms.

Maybe we can setup priorities automatically in most situations. We know sequences like: sys-new -> sys-firewall -> vmproxy1 -> sys-whonix -> vmvpn -> appvm 

So, we can stop them one-by-one from right to left. It is reasonable, then stop all of them at the same time.

Or maybe 2-3 groups of vms by priority, e.g:

default - automatic priority by sequences.

highest - must shutdown before any other if we can do that (no dependences) 

low - shutdown after all other

ps. as always sorry for my English

[Marek Marczykowski-Górecki]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

While there is some sense in having such sequential shutdown, it's
rather solution for this one, not the shutdown hung:
https://github.com/QubesOS/qubes-issues/issues/1826

If you execute `qvm-shutdown --all --wait` just before shutting down the
system, it will shutdown all the VMs at the same time, and it's also
enough to have fast shutdown.
The problem is that VMs shutdown during system shutdown (which
is called exactly as `qvm-shutdown --all --wait`) somehow doesn't finish
its task. VMs shutdown itself is initiated, which is visible for example
in disk activity. But I don't know what happens next - either not all
VMs are powered off, or some cleanup isn't executed. If you look at
shutdown messages, you can see at the end that disk can't be unmounted
because it is still in use. This may be some hint about the cause...

- --
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAEBCAAGBQJXcc7rAAoJENuP0xzK19csLEQIAIP+pxq7YpnddK2eIxpRt+QC
KREhBWEqBhKkw8x+mGvE4L2GPMKJ0xWsS9398s/XQ2gKcz24Ow79xfgWex+/pURS
bTqp7K7bVIW/sEHofO0/dLJSguNnx5r3X8TnwSXMkDNvdfQBm+6vm2R+zhwij5cb
Oiuchp6JMVkES16y67TVkPYm+Py9wcSxWj51hKgFqeDDcTveDmOCZiX+pBwDuXGN
Gei0AuZlq/wl9NmDnfyQFB2Db8zgg1MWRpY3LFT4eu7G+4IbQ0fKzghNPsoRH+SU
IS5ONxkFLlJzSGThhBHEeTuK9vvth2Ik551QI1X7f1Ve95OdzPyVB2BlerqWsLs=
=NeYG
-----END PGP SIGNATURE-----

[raah...@gmail.com]

I never had shutdown stall from not shutting down the vms that start at boot. Like all the sys vms. Its only happens sometimes if I don't shutdown an appvm first that I've manually started.

[Eva Star]

Maybe some hotlink for this operation at Qubes Manager menu ?