Increasing the number of concurrent streams while guaranteeing jitter-free operation is a primary issue for video servers. Disks storing popular videos tend to become overloaded, preventing the server accommodating more clients due to the unbalanced use of bandwidth. We propose an adaptive data retrieval scheme for load sharing in clustered video servers. We analyze how the data retrieval period affects the utilization of disk bandwidth and buffer space, and then develop a robust period management policy to satisfy the real-time requirements of video streams. We go on to propose a new data retrieval scheme in which the period can be dynamically adjusted so as to increase the disk bandwidth capacity of heavily loaded clusters and increase the number of clients admitted. Simulations demonstrate that our scheme is able to cope effectively with dynamically changing workloads and enables the server to admit many more clients.