NetworkManager-wait-online failing in lxc container since 1.28 update

Using the current daily Ubuntu image in lxc, the NetworkManager-wait-online service fails.

Journal from the container start log

Doing

# /usr/bin/nm-online -s 
gives a 
Connecting.......... <n>s
which countdowns for 30s then displays
Connecting...............    0s [startup-pending]

# nmcli device
DEVICE  TYPE      STATE      CONNECTION 
eth0    ethernet  unmanaged  --         
lo      loopback  unmanaged  --

new journal log with logging enabled

log

the log has

[1612348045.1487] manager: startup complete is waiting for device 'lo' (link-init)

changed the description

reverting 78dc57d8 fixes the issue @alfonsosanchezbeato

Updating after debugging a bit

• lxc unprevileged containers have udev enabled
• the commit mentioned earlier fixed the udev detection and makes n-m enable it
• the use of udev_enumerate_add_match_is_initialized makes udev enumeration not work under lxc

it could be a side effect of the recent systemd and udev changes mentioned on https://github.com/systemd/systemd/blob/main/NEWS#L13 , in any case not a network-manager so I'm closing the report. Thanks Thomas for the debugging hints

closed