systemd restart loop delays shutdown

[bcressey]

Image I'm using:
v0.11.7

Issue or Feature Request:
When testing locally in aws-dev, I issued an apiclient reboot command and hit an error where this pair of messages spammed the console during system shutdown:

[  135.289329] host-containers@admin[1168]: serial-getty@ttyS0.service has no holdoff time, scheduling restart.
[  135.290142] host-containers@admin[1168]: Stop job pending for unit, delaying automatic restart.

I've also heard scattered reports that enabling the admin container can cause delays when scaling in a cluster, and I suspect this might be the culprit.

systemd fixed this upstream in systemd/systemd@c972880 which we should add to the systemd patches here.