- Split playbooks into two, one for 3.0 minor upgrades and one for 3.0 to 3.1
  upgrades
- Move upgrade playbooks to common/openshift/cluster/upgrades from adhoc
- Added a byo wrapper playbooks to set the groups based on the byo
  conventions, other providers will need similar playbooks added eventually
- installer wrapper updates for refactored upgrade playbooks
  - call new 3.0 to 3.1 upgrade playbook
  - various fixes for edge cases I hit with a really old config laying
    around.
  - fix output of host facts to show connect_to value.

Instead of combining this with tasks to restart services, add a separate
started+enabled play for masters and nodes at the end of the playbook.

With the openshift to atomic-openshift renames, some services were not enabled
after upgrade. Added enabled directives to all service restart lines in the
upgrade playbook.

Remove upgrade playbook restriction on 3.0.2.

Rather than assuming the etcd data dir, we now read if from master-config.yaml
if using embedded etcd, otherwise from etcd.conf.

Doing so now required use of PyYAML to parse config file when gathering facts.

Fixed discrepancy with data_dir fact and openshift-enterprise deployment_type.

This is blocking 3.0.1 upgrades to 3.1 incorrectly, which is a scenario we
should support.

Skip some 3.1 checks if doing a 3.0.x to 3.0.2 upgrade.

Improve error message when oc whoami fails (i.e. openshift is down) during
pre-upgrade checks, rather than assuming the binary doesn't exist.