File system and file management method which realize distributed replication in system having shared type raid

摘要

When a file held on a shared type RAID (4) is to be updated, a file restoration data storage section (11) reads data before the update operation to hold it as file restoration data (14, 18) on both a primary computer (1) and a backup computer (2). When a checkpoint taking command is output from a checkpoint controller (5), this data is discarded. When a process (9) is aborted, a file restorator (12) restores the file to a state at the most recently taken checkpoint by using the file restoration data (14). When the primary computer (1) goes down, a file restorator (20) restores the file to a state at the most recently taken checkpoint by using the file restoration data (18). <IMAGE>