Avoid calling udevadm in containers

[jengelh]

Version: 5.0.3 (Tumbleweed 20240328 amd64)

Run yast-network in a LXC/nspawn container, set a new static hostname. When yast-network tries to save the configuration, it does that, but also runs into an error:

                               │ Execution of command "[["/usr/bin/udevadm", "control",       │                                
                               │ "--reload"]]" failed. Exit code: 1 Error output: Failed to   │                                
                               │ send reload request: No such file or directory               │                                

and

   -   Activate network service│ Error                                                        │                                
   -   Update configuration    │ Execution of command "[["/usr/bin/udevadm", "trigger",       │                                
                               │ "--subsystem-match=net", "--action=add"]]" failed. Exit      │                                
                               │ code: 1 Error output: host0: Failed to write 'add' to        │                                
                               │ '/sys/devices/virtual/net/host0/uevent': Read-only file      │                                
                               │ system                                                       │                                

This is because udev cannot be reloaded in a container.

# strace -f udevadm control --reload 2>&1|grep ENOE
connect(3, {sa_family=AF_UNIX, sun_path="/run/udev/control"}, 20) = -1 ENOENT (No such file or directory)

Expected:

Use systemd-detect-virt or similar to see if reloading even makes sense. Or perhaps there is a service that could be reloaded; if said service has ConditionVirt=, then something in the style of systemctl reload udevadm would ignore-succeed the request. (There is no udevadm.service though..)