Accidentally interupted Dom0 update. App-Qubes wont start

User Support
1

Last night I started to update Qubes to 4.2.3 using the GUI updater tool. Went to lock the screen while half paying attention. Not sure if I clicked the wrong thing, but the pc screen went black as if going into suspend. I woke it back up with the power button, as usual, and it looked like it was rebooting… Came back up and now I am unable to get any qubes to open at all, not even offline qubes. Getting this error:
Start failed: internal error: libxenlight failed to create new domain ‘qube-name-here’

I’m thinking I interrupted the update at a critical point and it screwed something up. Any advice? I’d rather not go through a system restore if I don’t have to.

Edit: but I suppose I should note that Dom0 comes up fine. The display manager is working properly.

2

Run this command in dom0 terminal:

journalctl -f -n0

Then try to start some qube e.g. vault.
Check the log messages and see if you’ll see some errors.

You can also check if dnf have some errors:

sudo dnf check

Maybe you can try to use sudo dnf history redo last or sudo dnf history undo last but I don’t know if it’ll work.

3

I ran journalctl -f -n0. The only relevant entries I see are:
vm.vault: Start failed: Cannot connect to qrexec agent for 60 seconds.
and maybe
WARNING: Thin pool qubes_dom0-vm–pool-tpool data is now 84.36% full.

In the /var/log/xen/console/guest-vault.log, I see:
PCI: Fatal: No config space access function found
ACPI Error: No handler or method for GPE 00, disabling event
acpi LNXCPU:03: Still not present [There are like 50 of these messages repeated]

The last message in the log is:
Waiting for /dev/xvdd device…

Running this I see 1 problem:
qubes-desktop-linux-menu-1.2.0-1.fc37.noarch is a duplicate with qubes-desktop-linux-menu-1.2.1-1.fc37.noarch

4

Maybe try to change the default linux kernel for the qubes in the Qubes Global Config → General Settings → Linux kernel to the older one.

5

That worked.
I switched from kernel 6.6.48-1.fc37 (the broken one) to 6.6.42-1.fc37, and now everything comes up normally.

After fixing the issue I applied the most recent dom0 update and tried switching the kernel back to 6.6.48-1.fc37, but that kernel still doesn’t allow me to open any qubes.

6

Maybe try to remove this kernel package and then update again.

sudo dnf remove kernel-qubes-vm-6.6.48-1.qubes.fc37
7

This fixed that particular kernel.
Thanks for the help!