DebAWeek: backupninja

“Backupninja is a silent flower blossom death strike to lost data.” That about sums it up. The premise of backupninja is fairly straightforward: allow packages and users to create backup jobs by putting a file in a directory called /etc/backup.d. These jobs rely on a handler script to execute and handler scripts can be customized to handle a particular type of backup job.

A database backup job for example might be to dump the contents of all databases to files suitable for backup by a typical file backup system. The program has several built in handlers for handling MySQL, PostgreSQL, LDAP, Maildir, Subversion, and Trac. It also has several methods for handling file backups including rdiff-backup, duplicity, and cd burning through mkisofs and cdrecord. It is also possible to extend backupninja by developing new handlers for other forms of backup.

This extensibility is one of the strengths of backupninja. Another benefit is the ability to schedule jobs easily using plain language such as “wednesday at 02:00”. However the program is definitely not a tool for novices and is perhaps best suited to server environments. There is a tool called ninjahelper that can aid in some of the tasks associated with backupninja but most administrators will probably find themselves configuring the job files in a text editor.

Though not very well known backupninja has a solid design that is extremely extensible. These traits can make it a great tool, if not as a sole backup solution, then as a valuable addition to a larger backup strategy.