PURPOSE: To provide high resposiveness even when processing jobs are expanded by always preventing a certain computer from being set in an idle state and effectively utilizing computer resources concerning the system for waiting operation mutually among plural computers so as to recover faults among computers one and another.
CONSTITUTION: In respect to plural computers CPU0-CPU3, the various kinds of job processing B0-B3 are separately performed by utilizing independent resources respectively stored in multi-call disks D0-D3, and the respective computers CPU0-CPU3 are operated as working systems to the job processing of the own systems and operated as standby systems to the job processing of the other systems. When the fault generation of a certain computer is informed through a shared memory SM, the computer at the succeeding destination is recognized based on succeeding information previously integrated in respective local disks C0-C3 as operation data, the job working resources related to the fault generation in the computer at the relevant succeeding destination is selectively read from the corresponding multi-call disk D through an MCP, and the fault job is succeeded.
TAKIMOTO HIDEAKI
IGUCHI MUNEKI
