Added zap_disk option

[cbaker423]

This change adds an attribute that enables the OSD disks to be zapped before activation. This is handy for HW that has already been formatted before use with Ceph, and requires a wipe before use with Ceph.

Also, this is my first commit to Ceph, so let me know if I did something inconsistent with the Ceph community contribution guidelines.

[clewis]

This is useful, but it makes me really nervous. This seems like an easy way to accidentally delete your entire cluster.

I'd say this should clear the zap_disk flag after it runs, much like ['ceph']['osd_devices'][index]['status'] is set to 'deployed'.

Maybe doing this at the individual OSD would be better? Maybe re-use the status field with a value of 'zap-disk' or something?

[cbaker423]

@clewis I agree; that is a good point to note. I thought about that a bit when creating the zap-disk option, but wondered if it would be too tedious to set the zap_disk option for each of the OSDs in the ceph environment. I guess the debate is between convenience and security.

For now, I will go ahead and try adding this attribute to each of the OSDs