
Refactored the inventory generation to allow for scaling up the openshift stack. They deployment process now checks whether the origin-node container is already running in the openshift nodes. All the nodes where origin-node is not present are considered as new. The new_masters/new_nodes sections will be populated accordingly and the `scaleup.yml` playbook will be used instead of the `deploy_cluster` one *only* if the stack is being updated. Unfortunately, openshift-ansible's implementation is not as idempotent as TripleO's, which requires differentiating between deployments, scaleup and between new and old nodes. Change-Id: I15c4e855a3ccfba6ce8d516b5b59e7508c4885cc
Externally Installed Services
The services in this directory and the Ansible hook they use (external_deploy_tasks) are currently considered experimental.