The redundancy of a file is maintained without using a centralized controller in a system where file servers are connected via a network. Each domain ND within a file storage system includes a file server FS, a monitoring server WS, and a client CL, which are connected via a LAN. When the file server FS detects a failure in another file server FS in the system, it considers that redundancy of a file stored in the failed file server FS has decreased, and then determines whether or not it is responsible for recovering the redundancy. If the file server FS has determined that it is responsible for recovering the redundancy, it selects a file server FS as a file copy source and a file server FS as a file copy destination, and then copies the identical file from the copy source server to the copy destination server.