I'd like to suggest, or at least open a discussion about, pushing cloudfs to F16 instead. It's not that it doesn't work, but I think it's insufficiently usable/useful as-is compared to how it could be in six months. My concern is that making it generally available now could create work and distractions that would only slow needed work, and also cause negative user experiences that would kill interest in what I still think is a fundamentally strong feature set. Specifically:

• The user experience is terrible, requiring far too many manual steps and far too much knowledge of how glusterfs already works. Partly this is a technical failure on my part, not figuring out the right integration points. Mostly, though, it's a matter of needing a far "thicker" layer of integration and administration tools than currently exist.

• The use cases for the multi-tenancy feature also more often than not mandate dynamic user (tenant) and management, which is not there yet.

• Stronger authentication and encryption are also must-haves for many potential users.

• Quota and billing support are explicitly supposed to be part of the feature, and are simply not ready.

• At least one user has mentioned that support for POSIX ACLs and/or SELinux would also make the result much more usable.

I understand that this puts more people than just me in a difficult position, and I apologize, but it seems that now is the time to ensure that neither the product nor Fedora's general reputation for quality are harmed. FWIW, I would consider a deferral an example of the Fedora process working - people thinking, talking, sharing ideas and deciding all before "going live" with something that just doesn't meet quality standards or advance our shared goals.