Different fields including biomedical-engineering, educational research and geology have an increasing need to process large amounts of video and make them electronically available at different locations. So far, this has been a failure-prone tedious operation with an operator needed to babysit the processing and off-site replication of processed video. In this work, we developed a fault-tolerant system that handles large scale processing and replication of digital video in a fully automated manner. The system is highly resilient and handles a variety of hardware, software and network failures making it possible to process videos using commodity clusters or grid resources. Finally, we discuss how the system is being used in educational research to process several hundred terabytes of video.