Current deployment explanation:
Deployment modifies single jupyter server deployment (bootstrap-jupyter) while keeping that deployment option intact. My thought behind this was the single server deployment option still provides the quickest and easiest route for someone just trying to demo what geowave + jupyter is capable of with the least additional configuration needed. JupyterHub with GEOAxis seemed a more specific deployment route that wouldn't traditionally be used by common public users.

For JupyterHub:
Current deployment assumes GEOAxis OAuthenticator will be used along with ssl certificates. These settings are configured within the jupyterhub_config.py. JupyterHub is configured as a Upstart service and started when the bootstrapping is complete. When JupyterHub starts it will run jupyterhub_config.py which loads ssl settings and sets up the user whitelist from files mentioned below.
Files used by the runtime configuration are expected to be located in the following directory:
/srv/jupyterhub/
Some examples of runtime config files would be the ssl certs and user list
/srv/jupyterhub/ssl/ssl.key
/srv/jupyterhub/ssl/ssl.cert
/srv/jupyterhub/userlist
/srv/jupyterhub/env_blacklist

Files used for configuring jupyterhub are expected to be in the following directory:
/etc/jupyterhub/
The jupyterhub config file will be located here, and the user is expected to put OAuth environment variables in a file here if they want the Upstart service to recognize them. /etc/jupyterhub/oauth_env.sh
When the Upstart service starts it will source conda + oauth dependencies to make sure they are available to the jupterhub daemon.

Questions about deployment:

1. Are we expecting to make this deployment compatible with other OAuth providers/Authenticators besides GEOAxis? i.e. github + PAM authentication?
2. If so, which OAuth providers/Authenticators should we support deployments for out of box?