We currently have checks for host.managed? and managed?in the dhp/dns/tftp orchestration, but they are not the first thing to be checked. as such, it's possible to call methods that don't exist for an STI host class with a new nic (eg discovery will attempt to call ip_available? on import, and fail)

We should move the managed checks to the beginning, since if a host/nic is not managed, it makes no sense to orchestrate anyway.