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"