Fail over method and a computing system having fail over function

A computing system which comprises a cluster constituted of a plurality of nodes and a system management node to manage the cluster, and in which, when a trouble occurs in one of the plurality of nodes, a processing of the troubled node is taken over to another node, comprising:

• a shared storage system which is shared by the nodes in the cluster and the system management node and which stores taking-over information when a trouble occurs in one of the plurality of nodes,
• wherein each node in the cluster comprises:
• a monitoring unit which monitors an operation state of another node;
• a detection unit which detects an occurrence of a trouble on the other node based on a result of the monitoring;
• a taking-over information obtaining unit which obtains the taking-over information from the shared storage system when the trouble is detected;
• and
• a fail over unit which takes over processings of the troubled node based on the taking-over information when the trouble occurs.

A computing system which comprises a cluster constituted of a plurality of nodes and a system management node to manage the cluster, and in which, when a trouble occurs in one of the plurality of nodes, a processing of the troubled node is taken over to another node, comprising:

• a shared storage system which is shared by the nodes in the cluster and the system management node and which stores taking-over information when a trouble occurs in one of the plurality of nodes,
• wherein each node in the cluster comprises:
• a monitoring unit which monitors an operation state of another node;
• a detection unit which detects one of an occurrence of a trouble and a recovery from a trouble on the other node based on a result of the monitoring;
• a taking-over information obtaining unit which obtains the taking-over information from the shared storage system when one of the trouble and the recovery is detected;
• a fail over unit which takes over processings of the troubled node based on the taking-over information when the trouble occurs;
• and
• a fail back unit which takes over the processings to the recovered node based on the taking-over information when the recovery is performed.