Three times in just the past week, I've had to help users on IRC who were trying to set up peer-to-peer style mounts between a group of servers, and made the "classic" mistake of using the same directories both as bricks and as client mountpoints.  In spirit similar to bug764139, I'd like to suggest that we try to detect and avoid at least the most obvious cases of this error.  One fairly simple trick would be to have storage/posix place an xattr (e.g. trusted.glusterfs.isabrick) on the root directory.  Then mount.glusterfs and nfs/server could check for presence of this xattr, and fail the mount/export if found.  I'm sure many other schemes could be used also.