Cluster controlling system operating on a plurality of computers in a cluster system

摘要

A cluster controlling system transfers packages which have been operating on one computer to another computer when a fault or failure has occurred by monitoring and controlling the packages in the entire system. When the respective packages are started-up, cluster daemons on the respective computers monitor and control resources on the operating computers. The monitored and controlled data are stored in the respective computers as local data. A manager communicates with cluster daemons on the respective computers, and stores data in a global data memory to monitor and control the entire system. The manager is actually one of the packages operating in the cluster system. If a fault or failure occurs in the manager or in the computer running the manager, the manager is re-started on another computer by a cluster daemon.