首页>
外国专利>
EXPANDING METHOD FOR CONTINUOUSLY MONITORING REMOTELY ACCESSIBLE RESOURCES IN ORDER TO PREPARE FOR NODE FAILURE IN WIDE CLUSTERS, ESPECIALLY RELATED TO A METHOD FOR MANAGING EXPANDED RESOURCES, WHICH IS USEFUL IN A SYSTEM WHICH HAS MANY NODES AND ARE TOLERABLE TO A MANAGEMENT NODE FAILURE FOR RESOURCES LOCATED IN REMOTE NODES
EXPANDING METHOD FOR CONTINUOUSLY MONITORING REMOTELY ACCESSIBLE RESOURCES IN ORDER TO PREPARE FOR NODE FAILURE IN WIDE CLUSTERS, ESPECIALLY RELATED TO A METHOD FOR MANAGING EXPANDED RESOURCES, WHICH IS USEFUL IN A SYSTEM WHICH HAS MANY NODES AND ARE TOLERABLE TO A MANAGEMENT NODE FAILURE FOR RESOURCES LOCATED IN REMOTE NODES
PURPOSE: An expanding method for continuously monitoring remotely accessible resources in order to prepare for a node failure in wide clusters is provided to continuously monitor and control remote resources without disconnection by transparently moving a proxy resource manager to other node, and to re-move the manager to an original node when a failed node is recovered, thereby realizing a load balancing. CONSTITUTION: A proxy resource manager is executed on plural nodes(110,120) of a distributed data processing system, and sequential resource generation identifiers associated with remotely accessible resources are stored within permanent storage devices connected to each of the nodes(110,120). The proxy resource manager is executed on at least other node, and states of the resources are maintained with a local version of the identifiers associated with the resources. The permanently stored identifiers are increased according to a restart of one of the nodes. According to a change of the resource state, the local identifier version is increased through a proxy resource agent. Local and permanent identifiers are compared together for the resources.
展开▼
