Description of problem:

When using rgw section in the ceph_conf_overrides it creates a ceph.conf in which the rgw section looks like:

For example :
~~~~
ceph_conf_overrides:
  rgw:
    rgw_op_thread_timeout: 900
~~~~

```
[rgw]
rgw_op_thread_timeout = 900
```

In the expected results, it should apply the configuration for all the RGWs in the cluster.

This change is exists in the upstream master branch :

https://github.com/ceph/ceph-ansible/blob/master/roles/ceph-config/templates/ceph.conf.j2

There's a request to backport this to RHCS 3