Networking broken in fresh Debian 13 (KDE) standalone, even after manual configuration

User Support
, , , ,
1

Qubes OS version: 4.3.0
Guest OS version: Debian 13 (Trixie) KDE
VM/qube type: Standalone HVM
Installation method: official iso for KDE version of Debian

Expected behavior: qube is able to connect normally after properly configuring its IP and gateway address in KDE Network Manager
Actual behavior: qube fails to have functional connectivity after multiple methods of configuring networking were tried

Details:
I installed Debian 13 (KDE) in a Qubes standalone HVM, but it is unable to connect, even after correctly setting modifying its “Wired connection” in KDE Network Manager, setting its IP address as 10.137.0.43, its netmask as 255.0.0.0, and its gateway as 10.137.0.8, which are all details shown in Qubes manager except the netmask, which Qubes manager shows as 255.255.255.255, but after looking into Qubes guides it’s for some reason recommended to set the netmask of standalone VMs to 255.255.255.0 instead. I tried both 255.255.255.0 and 255.0.0.0, as well as 255.255.255.255 as shown in Qubes manager, but none of these work and the connection is still non-functional, so I assume the netmask setting isn’t the problem.

AppVM qubes based on default Qubes templates, including the Debian template, connect fine over the gateway VM and require no additional configuration (because Qubes packages, such as qubes-core-agent-networking installed in these templates by default handle it automatically), so it shouldn’t be some kind of issue inherent to my Qubes setup itself.

I even tried manually installing that package and its dependencies by copying the package files over to the offline VM and installing them with apt in offline mode – however, this broke most of the VM’s functionality entirely, presumably due to misconfigured dependencies, which aren’t properly configured merely through installing them.

I’m at a loss for what else to try. Is anyone else able to reproduce the issue or figure out a way to make this work?

2

I cant reproduce this. I have standalones with network access and
exactly the configuration you cite.

Have you correctly set DNS resolvers?

What qube is at 10.137.0.8?

I never presume to speak for the Qubes team.
When I comment in the Forum I speak for myself.

3

Yes, I have. I tried both 10.139.1.1,10.139.1.2 as shown in qube manager and recommended in the guide, as well as the gateway VM’s address, but neither of them worked. DNS is probably not the issue here either, as the connection seems broken entirely.

sys-whonix, but I tried a custom gateway VM later, which didn’t work either.

What’s most bizarre is that after modifying Routes in KDE Network Manager (while trying the custom gateway VM instead of Whonix Gateway), ping 1.1.1.1 briefly started working after disabling and re-enabling the connection, but the connectivity died soon after and I struggled to make it work again, as the behavior was really inconsistent. Even while ping 1.1.1.1 was working, other traffic still did not, even after manually mapping target domains to destination IPs in /etc/hosts.

4

Update: I seem to only have this issue when connecting the standalone directly to Whonix Gateway or the custom gateway VM (based on Debian 13 minimal). When I (as a workaround) created a gateway VM based on the default Qubes Debian 13 Xfce template, connected that gateway VM to Whonix Gateway, and connected my standalone to the workaround gateway, suddenly everything works fine after proper configuration.
I still don’t know why Whonix Gateway and the custom gateway VM don’t work even while setting their addresses correctly, when AppVMs based on stock templates can also connect fine over them.