| 摘要 |
PROBLEM TO BE SOLVED: To provide a multi-server system enabling respective servers to recognize operation conditions of all servers in a system, switch processing in a short time at an occurrence of a failure, and prevent a system halt in order to solve a problem that a conventional multi-server system takes long time until all of the servers recognize the failure of the other server and cannot switch the processing promptly. SOLUTION: A duplicated manager and a plurality of agents are provided with a file for storing operational statuses of all the servers. When the agent detects the failure of the server, the file of the manager is updated and a change notice is transmitted to the manager. When a change request notice is received from the manager, the file of the manager is copied. When the manager detects the failure of the server, its own file is updated, and when the file is updated or when the change notice is received, the change request notice is transmitted to all of the operating agents. COPYRIGHT: (C)2006,JPO&NCIPI
|
