The first fully-automated nano-daemon designed to drive large-scale image data and document delivery systems
masterd ran from /usr/locl/bin and simply kept feeding large processing tasks to an automated document delivery system comprised of compute clusters, then susequently processed by a MySQL database-powered web server which provided analysis data in human-readable format. The daemon ran uninterrupted across distributed systems for over 10 years without modification or failure (and ridiculously low down time). The concept is too simple to fail!
The system was directly responsible for making data generated by life sciences systems available essentially forever, and the servers involved with heavy processing remained without user logins throughout their lifespans.