摘要 |
PROBLEM TO BE SOLVED: To enable a system reexecute a process without losing checkpoint information by dynamically switching a storage place to another computer other than a faulty computer if a fault has occurred to the computer, which stores checkpoint information on a process being executed by some computer. SOLUTION: If a fault has occurred to a computer which holds checkpoint information on a process being executed by some computer, a fault detection part 31 of a computer C1 detects the fault and informs a checkpoint information management part 32 of that. Then, the checkpoint information management part 32 refers to the contents of a saving place storage table 33 and selects a computer other than the computer which controls the current saving place. Consequently, the checkpoint information of processes can continuously be stored and is not lost, and the processes can be rolled back and reexecuted according to the nearest checkpoint information. |