HA-OS VirtualBox VM deadlocked after bad Zigbee2mqtt config

After my Debian supervised HA install stopped working (long story) I decided to start over with HA OS on a Virtualbox VM. It was all smooth sailing until I started setting up zigbee2mqtt. Previously I had used ZHA, but I heard that zigbee2mqtt could be better.

After arguing with the config options for a bit, I managed to get the correct serial path to my adapter. However, that seems to crash the zigbee2mqtt addon and everything else on the system. I thought that zigbee2mqtt would run in a container, but perhaps I’m wrong.

When it happens, I have to kill the VM via PID because even Virtualbox can’t get it to turn off. It stops responding to IP pings and the VM console doesn’t respond either. I have about 30 seconds after restart that I can use the HA console before the whole thing locks up.

Is this VM cooked, or is there a way to disable zigbee2mqtt with 30 seconds of typing?

After further testing, it seems that the issue is actually just from connecting the USB device to the VM. I created an entirely new Virtualbox VM following the same instructions, but this time connected the USB device of my Zigbee coordinator the first time I booted the VM.

It displayed the same behavior, which is the VM will get deadlocked after between 30 seconds and a minute. Sorry zigbee2mqtt for blaming you.

My guess is that something has changed with my Linux drivers (I updated the host OS since the last time tried running HA) and now my coordinator is not supported.