I have few question not directly related to the changes proposed in this pull request, but since the spec document is called Tendrl Ceph-installer Wrapper I would expect to see answers to the following questions here. If I'm wrong about that and this is not right place for such details, feel free to redirect me somewhere else:
How would tendrl user get to the ansible logs from ceph-installer? By that I mean stdout from ansible. Something like that is extremely useful for debugging. Would that require Tendrl API call?
Would it be possible to force stop the task? Eg. when ansible get stuck because of some error?
What would Tendrl do when ceph-installer crashes, fails or misbehaves in so other way?
Would it be possible to restart ceph-installer task later? Eg. I have a lab with 20 servers and start the task, but some were not provisioned correctly and ansible fails there. So I go and fix the problem on affected servers and now what?
How would you handle parallel execution so that when Tendrl runs ceph-ansible task, it will be executed on all storage machines at the same time? How would that affect design of flows and data model?
I have few question not directly related to the changes proposed in this pull request, but since the spec document is called Tendrl Ceph-installer Wrapper I would expect to see answers to the following questions here. If I'm wrong about that and this is not right place for such details, feel free to redirect me somewhere else: