NOTE: This project is effectively on ice, because I lost the practical use case that was the primary motivation for this project (this is 2009-07-16, though this has been the case for some time).

Current status: Initial design/implementation (not yet usable).

A free (MIT license) high-level front-end to tools such as rdiff-backup and duplicity (a "backup umbrella").

Tools such as rdiff-backup and duplicity are great in and of themselves, but when backing up several machines and several targets per machine, one immediately needs to begin scripting to support backup rotation, reporting, retention, and so on.

The intent is to create a single definitive tool for this purpose, to bring the "level of abstraction" for backups up a notch.

Goals include:

• Support for intelligent scheduling
• Configurable concurrency
• Dependencies between jobs
• Summary reporting
• Alert generation on failures
• Automatic log retention with easy inspection
• Built-in system integration (such as file system snapshotting when available)
• Support running periodically (trivially run from cron), or constantly as a daemon (more powerful, but requires tighter system integration)