FileUniq identifies duplicate files, and optionally removes them or replaces them with hard or soft links. It can use a persistent database. It is designed to work with heavily hard-linked files, such as those generated by cp -al / rsync backups.