It is known that the average query cost of a file in a multiple disk system can be dramatically reduced by storing redundant records of the file onto different disks. However, additional space is needed for the redundant records. In addition, the achievement of minimum redundance is difficult. Recently, Chen et al. (1994) showed that the problem of allocating all buckets of a multiple key hashing (MKH) file optimally among multiple disks for concurrent partial match retrieval can be reduced to a smaller sized subfile, called the remainder of the file. Based upon their reduction work, this article suggests two redundant storage schemes for MKH files in a multiple disk system. We point out that only modest copies of the remainder are necessary to be redundantly stored, in order to get perfectly optimal average query response time performance for an MKH file. We also emphasize that by using different clusterings for different copies,both the record organization cost and the bucket allocation cost can be simultaneously reduced, and thus, the cost can be dramatically reduced.