Title: Loss of network in Proxmox VE 7 upon reboot
Date: 2022-02-23 20:00

This morning I woke up to my hypervisor being down.
Because it's automatically rebooting upon kernel upgrades, maybe it was
crashing at boot time, the unblinking LED on the network card was also pointing
in this direction. So I plugged a screen in, and investigated: the machine was up,
and I got a shell.

```
root@pve:~# ip a | grep DOWN
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master vmbr0 state DOWN group default qlen 1000
3: iw0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default qlen 1000

# journalctl -e
Feb 23 13:30:25 systemd[1]: systemd-fsckd.service: Succeeded.
Feb 23 13:31:55 systemd[1]: ifupdown2-pre.service: Main process exited, code=exited, status=1/FAILURE
Feb 23 13:31:55 systemd[1]: ifupdown2-pre.service: Failed with result 'exit-code'.
Feb 23 13:31:55 systemd[1]: Failed to start Helper to synchronize boot up for ifupdown.
Feb 23 13:31:55 systemd[1]: Dependency failed for Network initialization.
Feb 23 13:31:55 systemd[1]: networking.service: Job networking.service/start failed with result 'dependency'.
Feb 23 13:31:55 systemd[1]: Reached target Network.
Feb 23 13:31:55 systemd[1]: Reached target Network is Online.
```

Looks like something is wrong in the `ifupdown2-pre` service, which is a
dependency of `networking`, timeout'ing after ~2 minutes. Let's take a look
to see what `ifupdown2-pre` is supposed to do:

```
# cat /usr/lib/systemd/system/ifupdown2-pre.service 
[Unit]
Description=Helper to synchronize boot up for ifupdown
DefaultDependencies=no
Wants=systemd-udevd.service
After=systemd-udev-trigger.service
Before=network.target

[Service]
Type=oneshot
TimeoutSec=180
RemainAfterExit=yes
EnvironmentFile=-/etc/default/networking
ExecStart=/bin/udevadm settle
```

What is `udevadm settle`?

```
# apropos udevadm 
udevadm (8)          - udev management tool
# udevadm --help | grep settle                                                                                                  
  settle        Wait for pending udev events
```

So I guess it's waiting on some events in a blocking manner, reaching the
timeout, and preventing the `networking` service from correctly being started.
Since my hypervisor doesn't have any fancy hardware relying on udev events
shenanigans, I simply <s>disabled</s> masked the service and rebooted:

```
# systemctl mask ifupdown2-pre.service 
# reboot
```

And things went back to normal. This is apparently a [more-or-less known
bug](https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=959985) since 2020,
sigh.