| 摘要 |
In a computer system having a cluster configuration, a reset command issued from each of computers to any of the other computers is transmitted to a reset control unit. A control module of the reset control unit judges whether a target of the newly inputted reset command conflicts with a target and a source of a reset command currently executed. If judging as no conflict, the newly inputted reset command is transmitted to a destination computer of the reset command, and then information of the transmitted reset command is stored as information of the reset command that is currently being executed. With this configuration, when one of the computers in which failure has occurred is reset by means of heartbeat mutual monitoring, it is possible to avoid a delay caused by a mutual reset or a repeated reset, and thereby to quickly reset the failed computer.
|
