PURPOSE: To reduce a load on a worker required for the recovery work of a fault by automatizing the selection of a spare medium and the decision of a recovery timing when the fault occurs in a medium.

CONSTITUTION: After the fault occurs in the medium, a processor 3 changes the medium name of the medium VOL1n to a dummy medium name, and following that, changes one medium VOL2t decided by the processor in the spare medium is changed to the medium name of the medium VOL1n in which the fault occurs. Next, data is copied and recovered from another medium with the same content as that of the medium VOL1n, and when recovery is completed, the processor 3 changes the state recognition of the medium VOL1n from a faulty medium to a normal medium, and following access to a file is performed in the same manner as the one for another medium. In such a way, a mistake can be eliminated in working without stopping a job when the fault of the medium is recovered and also, by reducing time and labor required for a recovery processing.