The ORA machinery gets gummed up when it gets very large submissions and
throws a lot of 500s.

Co-Authored-By: Jason Bau <jbau@stanford.edu>

Use the EDXAPP_LANG var which already exists.

Add github.com to ssh known_hosts for jenkins

fixing check for cluster config

Change the default cloudformation elb backend ports.

mongo_clustered defaults to !!null

Now they default to the same as what is set in ansible which should reduce the number of things
we have to override when running ansible on a VPC.

Update to use cluster enabling vars in both rabbit and mongo.

adding reconfigure option for

Move xqueue to supervisor

http://jenkins.edx.org:8080/view/Ansible/job/ansible-deploy

To ensure they are taken into account before we try to start the process
through supervisord.

supervisord as well as all the processes it controls are run as
`{{ common_web_user }}`. This allows us to connect to it as that
user which allows us to manage the processes as www-data, instead
of root.

Common refactor - Fix ora & xqueue

Feanil/mongo rs

Use the mongo path settings provided by the mongo role.

Jarv/xserver updates

Jarv/certs role

Since the data is not at a standard location anymore, set the location through
a environment variable, `NLTK_DATA`.