JP2008293441A - Method and apparatus for predicting device fault - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve problems that since the calculation or the like of correlation to predict the generation of a device fault is necessary, a prediction apparatus is complicated and whether a certain phenomenon is a previous phenomenon related to a fault can be obtained from the past data of fault analysis only in fault analysis after fault generation. <P>SOLUTION: If the existence of a phenomenon causing a fault before the generation of the fault is detected from a log at the time of fault generation when the fault is generated in a device 2, the phenomenon (fault pre-information) causing the fault is registered in a fault pre-information database 21 by an administrator terminal 1. When the fault pre-information is previously known, the fault pre-information is previously registered in the fault pre-information database 21, and when the fault pre-information is known thereafter, the fault pre-information is registered in the fault pre-information database 21 as required. In this embodiment, possibility of previously predicting a fault can be improved by storing the fault pre-phenomenon in the database. A fault prediction report part 22 reports fault prediction to an administrator or a user. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a device failure prediction method and a device failure prediction apparatus, and more particularly to a device failure prediction method and a device failure prediction device for predicting a failure occurring in a device in a field where the failure needs to be monitored.

  2. Description of the Related Art A performance monitoring device that predicts the occurrence of a device failure and monitors the performance of an information processing system is known (see, for example, Patent Document 1). FIG. 4 shows a configuration diagram of an example of the performance monitoring system described in Patent Document 1. In FIG. 1, the performance monitoring apparatus 10 includes a storage server 101 and an analysis server 102. The performance monitoring apparatus 10 is connected to an information processing system including a Web server 11, an AP (application) server 12, and a DB (database) server 13 through a communication line such as a local area network (LAN), and the communication line is connected to the information processing system. To monitor the status of each server.

  The accumulation server 101 accumulates, as monitoring data, transaction data indicating the throughput, processing name, and the like of transactions communicated through the communication lines connecting the servers 11 to 13 described above. Based on the monitoring data stored in the storage server 101, the analysis server 102 detects a failure that currently occurs in the information processing system or predicts a failure that may occur in the information processing system in the future.

  In this performance monitoring apparatus, the monitoring means (storage server 101) for monitoring the operation status of a plurality of information processing devices (servers 11 to 13) and the data communication status of each communication line connecting the plurality of information processing devices, This configuration includes a failure detection / prediction unit (analysis server 102) that detects a failure that currently occurs in the information processing system or predicts the possibility of a future failure in the information processing system based on the monitoring data. .

  Here, the failure detection / prediction unit performs information processing based on the correlation between the calculated types of monitoring data and the transition of the types of monitoring data obtained up to now by the monitoring unit. Predict that equipment may fail in the future. Further, the failure detection / prediction means uses the correlation between at least one of normal operation and abnormal operation of the information processing system calculated based on a plurality of types of monitoring data. Detect a failure that currently occurs, or predict the possibility that a failure will occur in the information processing system in the future.

JP 2005-327261 A

  However, the performance monitoring device described in Patent Document 1 that is a related invention of the device failure prediction device is configured to calculate a plurality of types of monitoring to predict that a failure may occur in the information processing device in the future. During normal operation of the information processing system calculated based on the correlation between data and the transition of multiple types of monitoring data obtained to date by the monitoring means, or based on multiple types of monitoring data Since the prediction is performed using at least one of the correlations during abnormal operation, the amount of calculation becomes enormous and the apparatus becomes complicated.

  To simplify the device, do not calculate correlations, do not predict all possible failures that may occur in the future, and detect that a phenomenon (prior phenomenon) that triggers the failure has occurred I can do it. However, whether or not a certain phenomenon is a prior phenomenon related to a failure can be obtained only from the past failure analysis data at the time of failure analysis after the occurrence of the failure.

  The present invention has been made in view of the above points, and by registering in advance a pre-phenomenon that triggers a failure according to whether or not the current phenomenon of the device is a pre-registered pre-phenomenon. An object of the present invention is to provide a device failure prediction method and device failure prediction apparatus that can predict a failure that will occur next.

  In order to achieve the above object, the first invention includes a first step of detecting a change in a system constituted by devices, and a failure prior phenomenon that occurs immediately before the detected change as a failure occurrence trigger. When the second step for determining whether or not the phenomenon is caused by the same phenomenon and the determination result obtained by the second step are caused by the same phenomenon as the failure prior phenomenon, the third step for notifying that effect is provided. And a step.

  In order to achieve the above object, the second invention includes a detecting means for detecting a change in a system constituted by devices, and a failure prior phenomenon that occurs immediately before the detected change as a failure occurrence trigger. A determination means for determining whether or not the phenomenon is caused by the same phenomenon, and a notification means for notifying that when the determination means obtains a determination result that the phenomenon is caused by the same phenomenon as the failure prior phenomenon. Features.

  According to the present invention, based on failure pre-phenomenon information that is a failure pre-occurrence phenomenon that occurs immediately before the occurrence of a device failure, predicting the occurrence of a device failure before the failure occurs and allowing the administrator to recognize it. Can do.

  Next, the best mode for carrying out the present invention will be described with reference to the drawings. FIG. 1 shows a block diagram of an embodiment of a device failure prediction apparatus according to the present invention. In the figure, an administrator terminal 1 is connected to a device 2. The device 2 includes a failure prior information database 21 and a failure prediction notification unit 22 in addition to each part (not shown) necessary for the original operation of the device. The prior failure information database 21 and the failure prediction notification unit 22 constitute a device failure prediction apparatus.

  The administrator terminal 1 is for registering a phenomenon that causes a failure in the failure prior information database 21. The administrator who operates the administrator terminal 1 may be a user who uses the device 2 or may be different from the user. In the failure prior information database 21, if a failure occurs in the device 2, if there is a phenomenon that triggers the failure before the failure occurs from the log at the time of failure occurrence, the phenomenon that triggers the failure (failure prior information ) Is registered (saved) by the administrator terminal 1. As for this prior failure information, all previously known information is registered in advance in the failure prior information database 21, and if found later, it is registered in the failure prior information database 21 as needed. In the present embodiment, the possibility that a failure can be predicted in advance increases by creating a database of failure prior phenomena.

  The failure prior information database 21 has not only a storage function but also a determination function. When the same phenomenon as any one of the registered failure prior information occurs in the device 2, the phenomenon Is predicted next, and notifies the failure prediction notification unit 22 of the predicted failure information. The failure prediction notification unit 22 notifies the administrator or user of the failure prediction by visual notification using a light emitting diode (LED) or the like, or by auditory notification using an alarm sound, and further by image display.

  The device 2 includes, for example, an ATCA switch based on a standard for advanced communication equipment (AdvancedTCA: Advanced Telecom Computing Architecture) established by PCI Industrial Computer Manufacturers Group (PICMG). (Port Link Down), and an example of failure prior information is port link up / down. That is, in the ATCA switch, when the port link up / down occurs, the port link may go down thereafter.

  Here, in a system including the administrator terminal 1 and the device 2, for example, as shown in FIG. 2A, a plurality of devices (here, three as an example) devices A, B, and C are managed as one device. As shown in FIG. 2B, there is a single device system 5 in which one device and one management device are integrated, as shown in FIG. Among these systems, in the system shown in FIG. 2 (A), the management device 4 that manages and manages the devices A, B, and C and the device of the single device system 5 shown in FIG. The apparatus failure prediction apparatus is composed of the prior information database 21 and the failure prediction notification unit 22.

  Next, the operation of the embodiment of FIG. 1 will be described with reference to the flowchart of one embodiment of the device failure prediction method of the present invention shown in FIG. Pre-failure information indicating a phenomenon that causes a failure is registered in advance from the administrator terminal 1 in the failure pre-information database 21 (step S1). Subsequently, the management device detects a change in the system shown in FIG. 2 (A) or (B) (step S2). The management device is not shown in FIG. This change in the system occurs when one or more devices constituting the system generate some phenomenon other than the original operation of the device.

  Subsequently, the failure prior information database 21 determines whether or not the detected change is the same phenomenon as the failure prior information registered in the failure prior information database 21 (step S3), and the detected change is detected. However, when the failure is the same phenomenon as the registered failure prior information, the failure prediction notification unit 22 is notified to that effect. As a result, the failure prediction notification unit 22 notifies the administrator of the occurrence of a phenomenon that triggers the failure, that is, the predicted failure (step S4).

  Here, when there are a plurality of failure prior information registered in the failure prior information database 21 and the failure prior information is identified and notified, the light blinking cycle, the long and short pattern of sound intermittently, Or, it is distinguished by a message displayed on the screen. Since these methods are publicly known, detailed description thereof is omitted. As a result, the administrator can recognize the possibility of a failure, and if the predicted failure affects the operation of the system, the relevant device is replaced.

  On the other hand, when it is determined in step S3 that the change detected in step S2 is a phenomenon that does not exist in the failure prior information registered in the failure prior information database 21, the phenomenon (system change) is detected. It is determined that this is not a phenomenon that triggers the occurrence of a failure, and no notification is sent from the failure prior information database 21 to the failure prediction notification unit 22 (step S5). Thereby, the failure prediction notification unit 22 does not perform any notification.

  As described above, according to the present embodiment, failure prior information indicating a phenomenon that triggers a failure is registered in the failure prior information database 21 in advance, and each time a phenomenon accompanied by a system change occurs, the phenomenon is detected in advance. Since it is determined whether or not it is registered in the information database 21, when it is determined that the phenomenon that has occurred is registered in the fault prior information database 21, the fault that occurs when the phenomenon occurs is referred to as the failure. You can predict and notify the administrator before you do. Further, according to the present embodiment, since it is not necessary to calculate a correlation and predict all failures that may occur in the future, the configuration of the device can be simplified.

  In addition, this invention is not limited to said Example, For example, not only when the administrator terminal 1 registers failure prior information in the failure prior information database 21, but automatically when a failure occurs A method of registering the failure occurrence trigger from the log in the failure prior information database 21 is also conceivable.

It is a block diagram of one Example of the apparatus failure prediction apparatus of this invention. It is a block diagram of each example of the system to which the equipment failure prediction apparatus of the present invention is applied. It is a flowchart of one Example of the apparatus failure prediction method of this invention. 1 is a configuration diagram of an example of a performance monitoring system described in Patent Literature 1. FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Administrator terminal 2 Apparatus 21 Failure prior information database 22 Failure prediction notification part

Claims (10)

A first step of detecting a change in a system constituted by devices;
A second step of determining whether or not the detected change is due to the same phenomenon as the failure prior phenomenon that occurs immediately before the occurrence of the failure;
A device failure prediction method comprising: a third step of notifying when the determination result that the failure is caused by the same phenomenon as the failure prior phenomenon is obtained in the second step. A method for predicting the occurrence of a failure of the device by an administrator who manages a system composed of one or a plurality of devices,
A first step of registering a phenomenon that occurs immediately before the occurrence of the failure in the database as failure prior information when the failure of the device occurs;
A second step of detecting a change in the system;
A third step of determining whether the detected change is due to the same phenomenon as the prior failure information registered in the database;
And a fourth step of notifying the administrator to that effect only when a determination result that the same phenomenon as the failure prior information is obtained by the third step is obtained. Failure prediction method.   The apparatus according to claim 2, further comprising a fifth step of registering, in the database, a phenomenon that occurs immediately before the occurrence of a failure when the failure occurs in the device. Failure prediction method.   The method further comprising: when a failure occurs in the device, automatically registering a phenomenon of a failure occurrence trigger from the log as the failure prior information in the database when the failure occurs. 2. The device failure prediction method according to 2.   A plurality of the prior failure information is registered in the database, and the fourth step indicates which failure prior information is the same among the plurality of previous failure information by light, sound or image. The device failure prediction method according to claim 2, wherein the determination results are distinguished and notified. Detecting means for detecting a change in a system constituted by devices;
Determining means for determining whether the detected change is due to the same phenomenon as the failure prior phenomenon that occurs immediately before the occurrence of the failure;
A device failure prediction apparatus, comprising: a notification means for notifying that when the determination means obtains a determination result that is due to the same phenomenon as the failure prior phenomenon. An apparatus for predicting the occurrence of a failure of the device by an administrator who manages a system composed of one or a plurality of devices,
A phenomenon that occurs immediately before the occurrence of the failure of the device as a trigger for occurrence of the failure is a database registered as failure prior information, and
Registration means for previously registering the prior failure information in the database;
Detecting means for detecting a change in the system;
Determining means for determining whether the detected change is due to the same phenomenon as the prior failure information registered in the database;
An apparatus failure prediction apparatus comprising: notification means for notifying the administrator only when the determination means obtains a determination result that is due to the same phenomenon as the prior failure information.   8. The apparatus failure prediction apparatus according to claim 7, further comprising means for registering, in the database, a phenomenon that occurs immediately before the occurrence of a failure when the failure occurs in the device. .   The apparatus according to claim 1, further comprising means for automatically registering, in the database, a failure occurrence trigger phenomenon from the log as the failure prior information when a failure occurs in the device. 8. The apparatus failure prediction apparatus according to 7.   A plurality of the prior failure information is registered in the database, and the notification means indicates, by light, sound, or image, which failure prior information is the same phenomenon among the plurality of previous failure information. The apparatus failure prediction apparatus according to any one of claims 7 to 9, wherein the result is distinguished and notified.

Images