Management method of computer system and management system
AssigneeHitachi, Ltd.
InventorsMasataka Nagura, Takayuki Nagai, Kiminori Sugauchi
Life of this patent
Estimating the term…
Abstract
To provide a fault analysis result classification function capable of reducing the time required to eliminate a fault in a monitoring target apparatus. The fault analysis results (fault cause candidates) obtained during the fault analysis processing are classified for the influence range of each of the fault cause candidates by analyzing which of the other fault cause candidates relate to the fault events of the apparatus abnormal state, which are used as the basis to derive the each of the fault cause candidates. Further, the classification results are displayed on a GUI.
Claims13
A management method of a computer system including one or more nodes to be monitored, and a management system which is coupled to the one or more nodes via a network and which monitors and manages the one or more nodes, wherein the management system performs the steps of:
- acquiring a processing performance value representing processing performance of the one or more nodes to detect occurrence of a fault in the one or more nodes on the basis of the acquired processing performance value,
- applying the detected fault to an analysis rule representing a relationship of a combination of one or more condition events capable of occurring in the one or more nodes with a conclusion event set as a fault cause of the combination of the condition events, to calculate a confidence value which is the information representing the possibility of occurrence of the fault in the one or more nodes,
- selecting, as a start-point cause candidate, one of the conclusion events set as a plurality of fault causes, to extract the condition events related to the start-point cause candidate,
- selecting, as a related cause candidate, a conclusion event which relates to the extracted condition event, and which is set as one or more fault causes different from the conclusion event selected as the start-point cause candidate, and
- performing classification processing of the conclusion event selected as the start-point cause candidate, and the conclusion event of the related cause candidate separately from the other conclusion events, to display the classified conclusion events on a display screen.
A management system coupled to one or more nodes to be monitored via a network and configured to manage the one or more nodes, comprising:
- a processor which acquires a processing performance value representing processing performance of the one or more nodes to detect the state of the one or more nodes on the basis of the acquired processing performance value;
- and
- a memory which stores an analysis rule representing a relationship of a combination of one or more condition events capable of occurring in the one or more nodes with a conclusion event set as a fault cause of the combination of the condition events,
- wherein the processor applies the detected state to the analysis rule to calculate a confidence value which is the information representing the possibility of occurrence of the fault in the one or more nodes;
- wherein the processor selects, as a start-point cause candidate, one of the conclusion events set as a plurality of fault causes, to extract the condition events related to the start-point cause candidate;
- wherein the processor selects, as a related cause candidate, a conclusion event which relates to the extracted condition events, and which is set as one or more fault causes different from the conclusion event of the start-point cause candidate;
- and
- wherein the processor performs classification processing of the conclusion event of the start-point cause candidate and the conclusion event of the related cause candidate separately from the other conclusion events, to display the classified conclusion events on a display screen.
Dependencies are inferred from the claim text.