PURPOSE: To enable all processing devices to continue to operate without being influenced by the occurrence of a fault in some processing device by detecting the occurrence of the fault in the shared memory of the individual processing device, and informing other processing device of it, and obtaining information stored in the shared memory of other processing device.

CONSTITUTION: For instance, in the case where the fault occurs in the shared memory 4 of the side of a magnetic tape control device 1, an error detecting part 22 detects it, and informs an access control part 5 of it, and a communica tion control part 6 informs the communication control part 6 of the side of the magnetic tape control device 1a of the occurrence of the fault. The access control part 5 of the side of the magnetic tape control device 1a sets an address and data at an input/ouput gate 7 to an input direction and an output direction respectively, and executes control operation so as to send the contents of the shared memory 4 designated by the address coming from the side of the mag netic tape control device 1 to the side of the magnetic tape control device 1. Thus, even if the fault occurs in some shared memory, all the control devices can be set always in an operating state by using other shared memory.