Bug 1625008

Name of function:
RHCS Ease of Use

User story:
As a administrator, I want to tune my Ceph cluster to take advantage of the information I already know about physical cluster so that I can… 

Purpose of function:
Improve deployment usability and make it more intuitive
This RFE Bugzilla has been created in order to track decisions and developments relating to a request for the ability to deploy second tier Ceph storage for OpenStack Platform, through OSP director.

Work required (from detail plan):
Procedural 
------
(example shows how this is installed)
Test: connect to your node and check your crush map (run a command on the overcloud) and see if it is what you expected.
How to tune my Ceph cluster 
How to set a custom crush map with director
------

Conceptual 
------
How is your Ceph cluster is physically laid out…  (geo-aware for redundancy in a different geo)
Must know how Ceph cluster is physically laid out (Why? Ceph is redundant. If you lose certain servers, you don’t lose data. With a big cluster, you can take advantage of geography - make cluster geo aware. “Replicate server somewhere else, geographically.” By telling Ceph to use a different geo for maximum safety, you make Ceph smarter; this is accomplished through the CRUSH map.
Info you need to tune your Ceph cluster
Ref: crush maps (bug comment 16)
What is a crush map?
Keep your templates - if you customize CRUSH map, make sure you include that file whenever you upgrade. In other words, when you use custom CRUSH maps, you enable this feature. If you disable the feature, you lose the customized maps.
------

Reference
------
N/A
------