Automate preparatory steps for running pipeline acceptance tests

Adding go-server auto register key.

New base box for devstacks that includes fixes for the chrome depreca…

Companion PR: https://github.com/edx/edx-analytics-pipeline/pull/204

Once this mv runs on the base box, it fails on provision

Just ensure the file is gone.

add credentials to /edx/bin/update

credentials: correct AWS_S3_CUSTOM_DOMAIN setting

Provide for Yarn/MapReduce memory allocation (Hadoop)

Ensure login shell is set correctly for Analytics users

Allow a single user access to multiple databases

allow for mapred/yarn resource config

use default mapreduce memory settings

corrected defaults for m1.medium, added yarn scheduler vars

testing with increased yarn am resource

remove task.io settings, defaults to AWS recommended

testing new hadoop conditional configuration

iter dict for conditional config

This avoids a class of confusing problems where a user might be given
access to two DBs and then when starting the app, fail because they can
only access one DB.

Quote username and password in migrate command

Otherwise the command fails if the password includes special shell
characters such as '&' or '<'

Define the CSMH database so it is created / migrate users defined

Workaround for debian squeeze deprecation and Chrome.

The older chrome repositories have been taken down; however, we
use a saved copy of the debian package. After installing that package,
ensure it is not referenced by the target machine after Chrome
is installed.

Remove remenants of removed vars.

Populate ENABLE_OAUTH2_PROVIDER in [cl]ms.env.yaml.

RABBITMQ_CLUSTERED was removed and rabbitmq_clustered_hosts was renamed.

Stop using the RABBITMQ_CLUSTERED var and tooling.

credentials: fix tags on requirements install task

It pulls the setting into the comment above it if you do that.

This method relied on the `hostvars` variable provided by ansible to always
be all the nodes in the rabbitmq cluster.  The hostvars var broke as a result
of the upgrade to ansible 1.9 but also even if it did work, this method
required that you always run rabbit plays on the whole cluster.  This is not
reasonable as it makes it very hard to heal a single node that may have fallen
over.

From here on, the RABBITMQ_CLUSTERED_HOSTS var should be used instead to
specify the rabbit hosts that are part of a cluster.

Small doc fix

Fix playbook name

Allow this to be run -c local so you don't need sudo or /usr/bin/python