What is the best practice for discovery of a clustered share/file system?