First of all, both the client and server will not be backwards compatible with older servers and clients. I've decided that because of the apparent privacy issues with the UUID, it would not be sane to support them. Therefore, the first feature is that there is now less code in both components.
Second of all, a file system engineer asked for some detailed reporting about file systems. I'm still refining what information is filtered out, but we will now store the following information about client machines:
- A list of all the file systems in use, and sizes
- The information that would be available from stat -f for each file system
- A mapping of mount points to file systems only when the mount point is already in the file system package in Fedora (or an equivalent list provided by the distro maintainer or administrator.)