Unfortunately, one of the containers locked up during restart
which looks like an issue with lxd.  (Ansible by default is talking
to 5 containers concurrently)

Also: make remove_dhcp optional

- roles for nagios
- build nagios-master from gold-master
- update hostname and postfix myhostname
- configure static IP address and /etc/hosts

Now the roles apply to the container itself, not to the host.
This simplifies the playbooks.

Also update dpkg check to work properly when python in state "un"

Create br-wan profile and apply it to gold-master, so it has networking