摘要 |
PROBLEM TO BE SOLVED: To improve availability by smoothly handing over the operation management of a cluster without using any wasteful stand-by node. SOLUTION: When a node 111 operating as a management node breaks down, a management node (node 115) is decided from among the other nodes 112 to 115. In the node 115, a node coordinator NC is started, and second cluster information 1400 is generated. The second cluster information 1400 is distributed to a node daemon ND of the node 115. In the node coordinator NC of the node 115, a new IP conversion table 1500 excluding the defective node 111 is prepared, and distributed to the nodes 112 to 115 excluding the defective node 111. First cluster information owned by the respective node daemon ND of the nodes 112 to 115 is updated to the second cluster information 1400, and an IP conversion table 1500 is handed over to each VNIC. COPYRIGHT: (C)2007,JPO&INPIT
|