In this paper, we present a novel disk failure recovery method that utilizes the inherent redundancy in video streams (rather than error-correcting codes) to ensure that the user-invoked on-the-fly failure recovery process does not impose any additional load on the disk array. We also present a disk array architecture that enhances the scalability of multimedia servers by: (1) integrating the recovery process with the decompression of video streams, and thereby distributing the reconstruction process across the clients; and (2) supporting graceful degradation in the quality of recovered images with increase in the number of disk failures.