Hey Ben,

Is this similar to the behavior that we were previously seeing when Docker wasn't properly allowing serial port access to the docker container? The fact that I've only seen this happen on our dockerized Linux bots seems suspicious.

Incidentally, that task you linked didn't run in docker. (Check the "inside_docker: 0" boolean in the bot's dimensions that ran that test.) So it's probably got something to do with the base linux. 

I recall there needing to be some system group modifications when deploying a battor to a linux box. Maybe that didn't happen/needs happening again?

Thanks for thinking of that Ben - SSHing into the bot, it looks like you're absolutely right. I think that the underlying bug here is that puppet should really be the one managing the dialout groups, and instead we're relying on manual intervention by the labs team. I'm going to go ahead and write a quick puppet class that can manage this for us.