JP2014021577A - Apparatus, system, method, and program for failure prediction - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a failure prediction device capable of predicting fault locations in an electronic device (hardware) in advance from execution time of a diagnosis program run on the electronic device.SOLUTION: A failure prediction device of the present invention includes: measurement means which measures execution time taken to diagnose each hardware constituent element of an electronic device in association with identification information of each hardware constituting element by a hardware constituent element unit; and reporting means which reports anomaly when execution time for any one of the hardware constituent elements exceeds a corresponding reference value set for the hardware constituent element.

Description

  The present invention relates to a failure prediction device, a failure prediction system, a failure prediction method, and a failure prediction program for detecting a failure inherent in the hardware of an electronic device without revealing it from the execution time of a diagnostic program for the electronic device About.

  Normally, maintenance of an electronic device is often diagnosed as a failure only when a failure occurs in the electronic device, and the cause of the failure is investigated by analyzing a log at the time of the failure, and is often dealt with by parts replacement or the like. However, when performing the above-mentioned maintenance response especially for an electronic device such as a magnetic disk, the system may not be able to start depending on the degree of failure, and important data may be lost. Therefore, in order to avoid such a problem, it is required to predict in advance a hardware failure of the electronic device before it becomes apparent.

  As a related technique for predicting a hardware failure of such an electronic device in advance, the Express Document 1 measures and stores the startup time from the power-on of the electronic device to the completion of the OS startup process, and the startup time. An abnormality inspection method for determining an abnormality when a predetermined time exceeds a predetermined time or when a deviation of accumulated startup time exceeds a predetermined range is disclosed.

  Patent Document 2 discloses an apparatus for measuring the execution time of a diagnostic program after powering on an electronic device and setting a time obtained by adding a predetermined margin to the time of the measurement result as a timeout time at the next system startup. Is published.

JP 2012-58782 JP JP 2005-165415 A

  In the techniques of Patent Document 1 and Patent Document 2 described above, abnormality detection is performed based on the startup time of the entire apparatus and the execution time of the diagnostic program. At this time, for example, even if the diagnosis execution time of any hardware component is long, if the diagnosis execution time of most other components is shorter than the reference value, the diagnosis execution time of the entire device As a result, there is a risk that the failure will be concealed. That is, the techniques of Patent Document 1 and Patent Document 2 cannot be expected to perform failure prediction with high accuracy.

  In addition, the techniques of Patent Document 1 and Patent Document 2 merely point out the possibility of hardware failure of the electronic device, and to identify which part of the hardware has the possibility of failure. Must wait for maintenance personnel to analyze the device log. If it takes time to identify a part where there is a possibility of failure and miss the timing of parts replacement, in the worst case, an actual failure may occur before parts replacement.

  In order to avoid the above-mentioned problem, it becomes a problem to perform failure prediction that can identify a portion having a possibility of failure with high accuracy.

  An object of the present invention is to provide a failure prediction device, a failure prediction system, a failure prediction method, and a failure prediction program that solve the above-described problems.

  The failure prediction apparatus according to an embodiment of the present invention measures the execution time required for diagnosis of a hardware component in association with identification information of the hardware component in units of hardware components included in the electronic device. And measuring means for reporting, in any one of the hardware components, a reporting means for reporting an abnormality when the execution time exceeds a reference value set in the hardware component.

  The failure prediction method according to an embodiment of the present invention measures the execution time required for diagnosis of the hardware component in association with the identification information of the hardware component in units of hardware components included in the electronic device. In any of the hardware components, when the execution time exceeds a reference value set in the hardware component, an abnormality is reported.

  The failure prediction program according to an embodiment of the present invention measures the execution time required for diagnosis of the hardware component in association with the identification information of the hardware component in units of hardware components included in the electronic device. And a reporting process for reporting an abnormality when the execution time exceeds a reference value set in the hardware component in any of the hardware components .

  The present invention makes it possible to predict in advance the location of an inherent failure without revealing it in the hardware of the electronic device from the diagnosis execution time of the electronic device.

It is a block diagram which shows the structure of the failure prediction system of 1st Embodiment of this invention. It is a flowchart which shows operation | movement of the 1st Embodiment of this invention. It is a structural example of the information stored in the reference | standard execution time memory | storage part in 1st Embodiment of this invention. It is a structural example of the information of the allowable time in 1st Embodiment of this invention. It is a block diagram which shows the structure of the failure prediction system of 2nd Embodiment of this invention.

  A first embodiment of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a block diagram showing the configuration of the failure prediction system of the present embodiment.

  The failure prediction system 1 of the present embodiment includes an electronic device 2 and a maintenance server 3. The maintenance server 3 is a server for inputting and outputting information for maintenance personnel of the electronic device 2 to perform maintenance work on the electronic device 2.

  The electronic device 2 includes a failure prediction device 10, a power input unit 20, a CPU unit 30, a memory unit 40, a disk unit 50, and a management unit 60.

  The power input unit 20 detects power-on to the electronic device 2 and transmits a power-on signal to the failure prediction device 10. The CPU unit 30, the memory unit 40, and the disk unit 50 each include a plurality of CPU modules, memory modules, and disk modules, and are replaced in units of modules when a failure occurs. The management unit 60 includes the log 61, manages the failure history of the electronic device 2, and reports the failure status to the maintenance server 3.

  The failure prediction apparatus 10 includes a POST (Power On Self Test) execution unit 11, a measurement unit 12, a configuration change detection unit 15, a reference execution time storage unit 16, and a report unit 17. The measurement unit 12 includes a POST execution detection unit 13 and a count unit 14.

  The POST execution unit 11 stores a POST program that is a diagnostic program for the electronic device 2 when the electronic device 2 is powered on. When the POST execution unit 11 receives a power-on signal from the power input unit 20, the POST execution unit 11 sequentially starts diagnosis of the CPU unit 30, the memory unit 40, and the disk unit 50.

  The POST execution detection unit 13 detects the start and end of POST execution for each module in the CPU unit 30, the memory unit 40, and the disk unit 50 by the POST execution unit 11. Then, the POST execution detection unit 13 transmits the detection result to the count unit 14 together with the module ID of the effective module of POST.

  When the count unit 14 receives from the POST execution detection unit 13 the start of POST execution and the module ID for which POST execution has started, the count unit 14 starts the count operation. When the count unit 14 receives the completion of POST execution of the module from the POST execution detection unit 13, the count unit 14 terminates the count operation and sends the count value together with the module ID of the module to the reference execution time storage unit 16 and the report unit 17. Send.

  The configuration change detection unit 15 detects the content of the configuration change when the power of the electronic device 2 is first turned on after the configuration change due to addition or deletion of a module in the CPU unit 30, the memory unit 40, or the disk unit 50 occurs. Then, the detection result is transmitted to the reference execution time storage unit 16.

  The reference execution time storage unit 16 stores the reference execution time of POST for each module of the CPU unit 30, the memory unit 40, and the disk unit 50, and new addition information of modules. A configuration example of the information stored in the reference execution time storage unit 16 is shown in FIG.

  In the example of FIG. 3, the CPU unit 30 includes four CPU modules whose module IDs are CPU0 to CPU3. The memory unit 40 includes eight memory modules having module IDs MEM0 to MEM7. The disk unit 50 includes four disk modules whose module IDs are DISK0 to DISK3.

  The reference execution time indicates a standard execution time of POST of each module. The reference execution time storage unit 16 stores the count value received from the count unit 14 as the reference execution time of the module when the power supply is first turned on after a new module is added.

  The newly added item is a flag indicating that a module has been newly added. The reference execution time storage unit 16 adds a record of the newly added module from the configuration change detection result received from the configuration change detection unit 15 and sets a value in the newly added flag of the record. In the example of FIG. 3, a module whose newly added flag is 1 is newly added, and the CPU3, MEM6, MEM7, and DISK3 modules are newly added. The reference execution time is not yet set for the newly added module, and the reference execution time storage unit 16 stores the reference execution time of the newly added module from the module ID and the count value received from the count unit 14.

  When the reporting unit 17 receives the module ID and the count value from the counting unit 14, the reporting unit 17 reads the reference execution time of the record in which the received module ID and the module ID match from the reference execution time storage unit 16. The reporting unit 17 adds the value of the allowable time 170 corresponding to the received module to the value of the reference execution time read from the reference execution time storage unit 16.

  The allowable time 170 is a count value and a reference execution time when the reporting unit 17 determines a module abnormality from the count value received from the count unit 14 and the reference execution time read from the reference execution time storage unit 16. It indicates the allowable value that the difference can take. A configuration example of the information of the allowable time 170 is shown in FIG.

  In the example of FIG. 4, the allowable time 170 includes values of 10 milliseconds, 50 milliseconds, and 2000 milliseconds as allowable times for the CPU unit 30, the memory unit 40, and the disk unit 50, respectively. Further, as shown in FIG. 3, the average reference execution times of the modules of the CPU unit 30, the memory unit 40, and the disk unit 50 are about 20 milliseconds, 100 milliseconds, and 4000 milliseconds, respectively. Therefore, in the present embodiment, the allowable time 170 is a value that is approximately half of the standard reference execution time, but this is an example, and the value of the allowable time 170 may be set based on a different reference. Note that the value of the allowable time 170 may be set manually by the administrator of the failure prediction system 1 or may be standardized from the reference execution time of each module stored in the reference execution time storage unit 16 by the failure prediction apparatus 10. In some cases, a basic reference execution time is calculated and set by a predetermined calculation formula based on the calculated value.

  The reporting unit 17 compares the added value of the reference execution time and the allowable time of each module described above with the count value received from the counting unit 14. When the count value is larger, the reporting unit 17 determines that the diagnosis target module is abnormal, and reports the occurrence of the abnormality to the management unit 60 together with the module ID of the module.

  For example, in the case of the module of CPU0, the reference execution time is 20 milliseconds as shown in FIG. 3, and the allowable time of the CPU unit is 10 milliseconds as shown in FIG. Therefore, when the count value related to CPU0 received from the count unit 14 is larger than 30 milliseconds obtained by adding the above two values, the report unit 17 reports the module abnormality of the CPU0 to the management unit 60.

  Next, the operation of this embodiment will be described in detail with reference to the flowchart of FIG.

  The POST execution unit 11 receives a power-on signal from the power input unit 20 (S101). The POST execution unit 11 starts diagnosis for the CPU unit 30, the memory unit 40, and the disk unit 50 in order (S102).

  The POST execution detection unit 13 detects the start of diagnosis by the POST execution unit 11, and transmits it to the count unit 14 together with the module ID of the module for which the POST execution unit 11 has started diagnosis (S103). The count unit 14 starts a count operation (S104). The POST execution detection unit 13 detects the end of diagnosis of the module by the POST execution unit 11 and transmits it to the count unit 14 (S105). The count unit 14 stops the count operation, and transmits the count value and the module ID of the module to the reference execution time storage unit 16 and the report unit 17 (S106).

  The reference execution time storage unit 16 checks the new addition flag of the record whose module ID matches the module ID of the module received from the counting unit 14 (S107). When the value of the new addition flag is 1 (Yes in S108), the reference execution time storage unit 16 writes the count value received from the count unit 14 in the item of reference execution time of the record corresponding to the module (S109). ). If the value of the new addition flag is 0 (No in S108), the process proceeds to S110.

  The reporting unit 17 reads the reference execution time value of the record corresponding to the module in the reference execution time storage unit 16 and adds the allowable time value of the unit to which the module belongs described in the allowable time 170. Then, the reporting unit 17 compares the added value with the count value received from the counting unit 14 (S110). When the count value is larger than the above addition value (Yes in S111), the reporting unit 17 reports the abnormality of the module together with the module ID to the management unit 60 (S112). If the count value is less than or equal to the above addition value (No in S111), the process proceeds to S113. If the module is not the last diagnostic module by the POST execution unit 11 (No in S113), the process returns to S103. When the module is the last diagnostic module by the POST execution unit 11 (Yes in S113), the entire process ends.

  In the present embodiment, there is an effect of predicting in advance a place where there is a possibility of failure in the hardware of the electronic device. The reason is that the reporting unit 17 compares the POST execution time for each module counted by the counting unit 14 with the POST execution time that is the reference of the module stored in the reference execution time storage unit 16 and counts This is because if there is a module whose result is longer than the predetermined allowable value, it is reported that the module is abnormal.

  The hardware of the electronic device may continue to repeat a retry operation in a data read / write operation for a while before becoming completely in a failed state. In this period, a retry operation occurs even when POST is executed, and therefore the POST execution time becomes longer than normal. Therefore, a hardware failure can be predicted in advance by comparing the POST execution time with the normal POST execution time.

In this embodiment, since the POST execution time is compared in units of hardware modules, the failure prediction is performed with higher accuracy than the conventional method in which the POST execution time is compared in the entire electronic device. Is possible. Then, it is possible to specify in advance a module whose POST execution time is equal to or greater than the allowable value as compared with the normal time as a module with a possibility of failure. As a result, it is possible to avoid a system failure by quickly replacing the module before the module actually fails.
<Second Embodiment>
Next, a second embodiment of the present invention will be described in detail with reference to the drawings.

  FIG. 5 is a block diagram showing the configuration of the failure prediction system according to the second embodiment of the present invention.

  The failure prediction system 1 includes a failure prediction device 10. The failure prediction apparatus 10 includes a measurement unit 12 and a report unit 17. The measurement unit 12 determines the POST execution time in units of hardware modules based on a diagnostic program for hardware components of the electronic device, for example, execution information from a POST execution unit that executes POST. Measured in association with information for identifying.

  The reporting unit 16 reports an abnormality in any hardware module when the POST execution time exceeds a reference value set in the module.

  In the present embodiment, as in the first embodiment, there is an effect of predicting in advance with high accuracy a place where there is a possibility of failure in the hardware of the electronic device. The reason is that the reporting unit 17 reports an abnormality when there is a module whose POST execution time for each module measured by the measuring unit 12 is longer than the reference value.

  In this embodiment, a value serving as the base of the above-described reference value may be stored in advance in a storage unit such as the reference execution time storage unit 16 in the first embodiment, or a failure may occur during POST execution. The user of the prediction system 1 may input manually.

  While the present invention has been described with reference to the embodiments, the present invention is not limited to the above embodiments. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

DESCRIPTION OF SYMBOLS 1 Failure prediction system 2 Electronic apparatus 3 Maintenance server 10 Failure prediction apparatus 11 POST execution part 12 Measurement part 13 POST execution detection part 14 Count part 15 Configuration change detection part 16 Reference | standard execution time memory | storage part 17 Report part 170 Permissible time 20 Power supply input part 30 CPU part 40 Memory part 50 Disk part 60 Management part 61 Log

Claims (10)

Measuring means for measuring the execution time required for diagnosis of the hardware component in association with identification information of the hardware component in units of hardware components included in the electronic device;
In any one of the hardware components, a reporting means for reporting an abnormality when the execution time exceeds a reference value set in the hardware component;
A failure prediction apparatus comprising: Detecting means for detecting a configuration change of the hardware component;
The failure according to claim 1, wherein the reporting unit sets, as the reference value, a value obtained by adding a predetermined allowable value to the execution time of the hardware component when the detection unit detects the configuration change. Prediction device. The failure prediction apparatus according to claim 1, wherein the reporting unit reports the identification information of the hardware component that has detected the abnormality.   A failure prediction system including the failure prediction device according to claim 1, the electronic device, and a maintenance server connected to the electronic device. For each hardware component included in the electronic device, the execution time required for diagnosis of the hardware component is measured in association with the identification information of the hardware component,
A failure prediction method for reporting an abnormality in any one of the hardware components when the execution time exceeds a reference value set in the hardware component. Detecting a configuration change of the hardware component;
The failure prediction method according to claim 5, wherein a value obtained by adding a predetermined allowable value to the execution time of the hardware component when the configuration change is detected is set as the reference value. The failure prediction method according to claim 5, wherein the identification information of the hardware component that has detected the abnormality is reported. Measurement processing for measuring the execution time required for diagnosis of the hardware component in association with the identification information of the hardware component in units of hardware components included in the electronic device;
In any one of the hardware components, a report process for reporting an abnormality when the execution time exceeds a reference value set in the hardware component;
Failure prediction program that causes a computer to execute. Causing a computer to execute a detection process for detecting a configuration change of the hardware component;
The failure of claim 8, wherein the reporting process sets, as the reference value, a value obtained by adding a predetermined allowable value to the execution time of the hardware component when the detection process detects the configuration change. Prediction program. The failure prediction program according to claim 8, wherein the reporting process reports the identification information of the hardware component that has detected the abnormality.

Images