The major goal of this patchset is to consolidate all machine types and their
corresponding facts logic in one place.  As more machine types were getting
created it became unrealistic to have all facts on all machine types.  There
were also quite a few complexities involved in the ordering of regexes and such
which made the whole process very error prone.  There is now a config file
dropped in /etc/everest by puppet that houses all the logic.  The external
nodes script, the nodetypes cgi, and everest-bootstrap all use that file as the
canonical source of information.

Another goal of this patchset is to have all the various responsibilities of
everest-bootstrap at the same level of abstraction. The way we were using the
'main' gem was a little out of hand.  In one block you had configuration,
bootstrap logic and two different DSLs.  That was a little absurd.  This will
allow us to test the tool alot more.  I even committed a vew trivial unit tests
for the DDNS stuff.

Instead of adding more and more complex command line options I'm taking the
'wizard' approach.  This should help to educate people and over time we can
make this a little more non-interactive.

Tickets fixed:
    Making the regex used for scraping the DDNS hash a little more strict
    https://engineering.redhat.com/trac/IT_Architecture/ticket/71
    Added some tests

    Removing CGI call from everest-bootstrap
    https://engineering.redhat.com/trac/IT_Architecture/ticket/40

    Host bootstrapping will now prompt the user to finish
    https://engineering.redhat.com/trac/IT_Architecture/ticket/45

    everest-bootstrap will now forcefully unmout when cleaning up
    https://engineering.redhat.com/trac/IT_Architecture/ticket/39

    everest-bootstrap now requires '--repo'
    https://engineering.redhat.com/trac/IT_Architecture/ticket/77

    Improving everest-bootstrap DDNS error message
    https://engineering.redhat.com/trac/IT_Architecture/ticket/65