We have to keep upstart from killing supervisor.

Supervisor could wait up to 5 days for workers to finish jobs. So
Upstart needs to wait that long for supervisor to finish.

Increase how long supervisor waits for workers to stop.

This is set as 5 days for now because we have seen "grade_course" tasks that
take about 4 days to complete for large classes and if we shutdown before
they are done we end up in a bad state in the database.

Update the course ID for the demo course

fixing perm issue with npm registry

adm

allows overriding ssh user

Make NAT monitor only warn about NAT failures.

removing internal network allow for basic auth

Adding force option for installing mongo

Log the lenght of all the queues.

refactoring to support multiple forward targets

no-dash

Adding prefix to namespace

path fixes

gitignore OSX files.

set basename the same way we do for ansible-private

adding extra config tasks

Feanil/update create dbs

Master ... release