JPH01177642A - System for diagnosing fault in computer system - Google Patents

Abstract

PURPOSE:To increase probability to confirm a defective point by activating an optimum diagnostic program based on the analysis result of fault information, executing a fault diagnosis, and repeatedly executing it prescribed times at a fixed time interval. CONSTITUTION:When a hard fault is generated in a computer 1 main body, fault information 10 is analized by a fault information analyzing part 7, and optimum micro diagnostic program selection information 11 is sent to the computer 1 main body. Based on the micro diagnostic program selection information 11, the optimum diagnostic program to the fault information 10 in a micro diagnostic program 4 stored in a control storage 2 is activated, and the fault is automatically diagnosed by the diagnostic program. By a timer 8 and a counter 9, the micro diagnostic program selection information 11 is sent to the computer 1 main body at the fixed time interval, and the fault is automatically diagnosed. Thus, even at the time of an intermittent fault, the probability to confirm the defective point can be increased.

Description

[Detailed Description of the Invention] [Summary] This invention relates to a fault diagnosis method for a computer system that automatically diagnoses the fault accurately and quickly when a hardware fault occurs during the operation of the computer system. , only when a hardware failure occurs, starts the most suitable diagnostic program among the microdiagnostic programs stored in the control storage to automatically perform failure diagnosis, and repeats the execution a predetermined number of times at regular time intervals to identify the defective location. The fault information generated when a hardware fault occurs is analyzed by the fault information analysis section, and based on the analysis results, the most suitable diagnosis is selected from among the micro-diagnosis programs stored in the control storage. Start the program and perform fault diagnosis,
By repeating this a predetermined number of times at regular time intervals, the possibility of confirming the defective location is increased.

[Industrial application field]

The present invention relates to a fault diagnosis method for a computer system that automatically diagnoses a defective location accurately and quickly when a hardware fault occurs during operation of the computer system.

[Conventional technology]

When operating a computer system, the reliability of the computer system itself is an extremely important issue. In order to improve this reliability, it goes without saying that it is necessary to improve the reliability of the hardware, but when a failure occurs, it is important to accurately diagnose the location of the failure and the type of failure as quickly as possible and perform maintenance. is important.

Conventionally, as shown in FIG. 3, a microprogram stored in a control storage (C3) 102 built in the computer main body 101 and configured with RAM, etc., and a service processor 1 provided outside the main body 101 are known.
A computer system equipped with 03 is known.

In this computer system, when a hardware failure occurs on the computer main body 101 side, failure information 104 is notified to the service processor 103 and logged there. The computer system is not configured to perform fault diagnosis, and the fault diagnosis will be performed by maintenance personnel after this.

Note that in the above system as well, a diagnosis is temporarily performed by a microprogram using the control storage 102 at the start of operation, but no diagnosis is performed during operation.

In recent years, more than 80% of hardware failures in computer systems are intermittent failures, and in many cases the system has returned to normal operating condition by the time maintenance personnel arrive at the site. For this reason, diagnosis of the defective location after a certain period of time has elapsed from the occurrence of the failure must be estimated from the logging information accumulated in the service processor 103, making it difficult to obtain accurate confirmation of the defective location.

Furthermore, in this case, there are many cases in which no abnormality is found even if the defective location estimated based on the logging information accumulated in the service processor 103 is brought back for detailed investigation.

[Problem that the invention seeks to solve]

In the conventional computer system described above, when a hardware failure occurs in the computer itself, failure information is notified to the service processor and logged there, but this is not a configuration that automatically executes failure diagnosis. Since the diagnosis is performed by maintenance personnel based on logging information, there is a problem in that it is difficult to confirm the location of the fault, especially in the case of intermittent faults.

The present invention is capable of automatically executing a fault diagnosis by activating an optimal diagnostic program among micro-diagnosis programs stored in a control storage only when a hardware fault occurs, and repeating the execution a predetermined number of times at fixed time intervals. The purpose of this invention is to provide a method for diagnosing faults in computer systems that can accurately identify defective locations.

[Means for solving problems]

As shown in the principle diagram of FIG. 1, a computer main body 1 has a control storage 2 in which a microprogram 3 that executes normal processing and a plurality of microdiagnosis programs 4 that execute fault diagnosis processing only when a hardware failure occurs are stored. and a logging unit 6 that stores failure information 10 generated when a hardware failure occurs on the computer main body 1 side, and analyzes the logging information from this logging unit 6 to obtain optimal micro-diagnosis program selection information 11. a fault information analysis section 7 that sends it to the computer main body 1 and starts the micro-diagnosis program 4; a timer 8 that repeatedly sends the micro-diagnosis program selection information @11 to the computer main body 1 after a certain period of time has elapsed; Program selection information 11
and a service processor 5 having a counter 9 for executing fault diagnosis by an optimal micro-diagnosis program 4 based on a predetermined number of times.
0 is logged by the logging unit 6 of the service processor 5, and this logging information is analyzed by the failure information analysis unit 7 to obtain optimal micro-diagnosis program selection information 11.
is sent to the computer main body 1, the optimum microdiagnosis program selected based on the microdiagnosis program selection information 11 is activated, and the timer 8 and counter 9 perform fault diagnosis using the microdiagnosis program a predetermined number of times at fixed time intervals. It is designed to be executed automatically.

[For production]

When a hardware failure occurs in the computer main body 1, this situation is notified to the service processor 5 as failure information 10. For service processor 5, the above failure information @10
is logged by the logging unit 6, and this logging information is analyzed by the failure information analysis unit 7, from which optimal micro-diagnosis program selection information 11, which is the analysis result for the failure information 10, is sent to the computer main body 1. .

Based on the micro-diagnostic program selection information 11, the optimal diagnostic program for the fault information 10 among the micro-diagnostic programs 4 stored in the control storage 2 is started, and the fault is automatically diagnosed by this diagnostic program. is executed.

At this time, the timer 8 and counter 9 provided in the service processor 5 cause the failure information analysis unit 7 to
The micro-diagnosis program selection information 11 from the computer is sent to the computer main body 1 at fixed time intervals, and fault diagnosis using the optimal micro-diagnosis program is automatically executed a predetermined number of times, so even intermittent faults can be detected. There is a higher possibility of confirming the location.

〔Example〕

An embodiment according to the present invention will be described below.

FIG. 2 is a block diagram of a fault diagnosis method in a computer system that is an embodiment of the present invention.

When a hardware failure occurs in the computer main body 1, the failure is detected by the hardware failure detection unit 12, and from there the failure information 10 indicating that a hardware failure has occurred is notified to the logging unit 6 of the service processor 5.

In the service processor 5, the fault information 10 is logged by the logging unit 6, and at the same time, the fault information analysis unit 7 analyzes the defective location based on the logging information. 11 is a micro diagnostic program control unit 1 of the computer main body 1;
Sent on 3rd. Then, based on the microdiagnosis program selection information 11 from the fault information analysis section 7, the microdiagnosis program control section 13 starts the most suitable diagnosis program from among the microdiagnosis programs 4A, 4B, -'-4N. , automatically executes fault diagnosis for the fault information 10.

As described above, the failure information analysis section 7 analyzes the defective location based on the logging information from the logging section 6, and sends the optimal microdiagnosis program selection information 11, which is the analysis result, to the microdiagnosis program control section 13. However, at this time, the failure information analysis section 7 analyzes the transmission frames 1, 2, . . . . The micro-diagnosis program to be activated is designated by the bit placed in the N frame position, and this is sent out as the micro-diagnosis program selection information 11.

Therefore, when the bit indicating the analysis result in the fault information analysis section 7 stands at the 1 position of the transmission frame 7A, the micro-diagnosis program control section 13 that receives this starts the micro-diagnosis program (1) 4A. Similarly, when the bit indicating the analysis result stands at the N position of the transmission frame 7A, the microdiagnosis program control unit 13 starts the microdiagnosis program (N) 4N.

The above-mentioned micro-diagnosis program is subdivided so that diagnosis is performed for each block on the hardware configuration of the computer main body 1. For example, the micro-diagnosis program (1
) 4A is the calculation block, micro diagnostic program (2)
4B is a control block and a micro diagnostic program (
N) 4N is a diagnostic program for the drive and reception blocks.

Now, if the bit indicating the analysis result in the failure information analysis unit 7 based on the failure information 10 from the hardware failure detection unit 12 stands at position 2 of the transmission frame 7A, the microdiagnosis program control unit 13 executes the microdiagnosis program. (
2) 4B is activated, and in this case, fault diagnosis in the control block of the computer main body 1 is executed.

As mentioned above, when a hardware failure occurs in the computer main body l, the failure information 1 is sent from the hardware failure detection unit 12 to the logging unit 6.
0 is notified, and the analysis result by the failure information analysis unit 7 based on this logging information is sent to the microdiagnosis program control unit 13 as the microdiagnosis program selection information 11.
The optimum micro-diagnosis program for the fault information 10 is activated and fault diagnosis is automatically executed.

Further, the service processor 5 uses the timer 8 to set the micro diagnostic program selection information 1 after a certain period of time has elapsed.
1 to the micro-diagnosis program control section 13 of the computer main body 1
The micro-diagnosis program is sent again to execute a fault diagnosis based on the above-mentioned optimal micro-diagnosis program, and this is repeatedly executed a predetermined number of times set by the counter 9.

In the above-mentioned fault diagnosis, if the fault location is confirmed in one fault diagnosis based on the micro-diagnosis program selection information 11 from the fault information analysis unit 7, the fault is detected by the micro-diagnosis program at that point. Finish the diagnosis. The diagnosis result 16 is then notified from the diagnosis result sending unit 14 to the logging unit 15 of the service processor 5, and the logging unit 15 logs the diagnosis result 16 and displays it on a display device or the like as appropriate.

Even if it is not possible to confirm the location of the defect with the above-mentioned one-time fault diagnosis, the fault diagnosis using the optimal micro-diagnosis program is repeated a predetermined number of times after a hardware fault occurs, so it is not possible to confirm the location of the fault in the case of an intermittent fault. However, there is a greater possibility of confirming the location of the defect. Furthermore, the micro-diagnosis program described above has a short execution time and a high failure detection rate for hardware failures. However, especially in the case of intermittent failures, it may not be possible to confirm the location of the failure even if the failure diagnosis is repeated a predetermined number of times. The information is notified to the logging unit 1-5 of the computer, and the information is logged here and a message to that effect is displayed as appropriate.

Note that the above-mentioned failure information analysis unit 7 sends the micro-diagnosis program selection information 11 to the micro-diagnosis program control unit 13 if the hardware failure is serious and the computer main body 1 has stopped operating due to it. do not have to.

In addition, the microprogram 3 stored in the control storage 2 normally executes normal processing that is not based on the microdiagnosis program 4 described above, and this normal processing is performed integrally with the hardware of the computer main body 1. We are creating specifications.

The normal processing described above guarantees the operation of the operating system (O8), and in this normal processing, the processing speed is faster if the execution steps per time are as short as possible. Since the fault diagnosis by the micro-diagnosis program 4 is executed only when a fault occurs, the influence of this fault diagnosis on the processing speed can be reduced as much as possible.

As mentioned above, when a hardware failure occurs, it is possible to start the most suitable diagnostic program among the microdiagnostic programs, and it is also possible to automatically execute failure diagnosis by repeating it a predetermined number of times at regular intervals using a timer and counter. This increases the possibility of confirming the location of the defect, and reduces the impact of low performance of the operating system as much as possible.

〔Effect of the invention〕

According to the present invention as described above, only when a hardware failure occurs, the most suitable diagnostic program among the microdiagnostic programs stored in the control storage is activated, and the failure diagnosis is automatically repeated a predetermined number of times at regular time intervals. This configuration allows hardware failures to be diagnosed using the most suitable diagnostic program, and since failure diagnosis is executed only when a hardware failure occurs, it is possible to minimize the impact on the operating system's low performance. A special effect can be obtained, such as increasing the possibility of confirming the location.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the principle of the present invention, FIG. 2 is a block diagram showing an embodiment of the present invention, and FIG. 3 is a block diagram showing a conventional example. ■...Computer body, 2...Control storage, 3...Micro program, 4A, 4B...4N...
...Micro diagnosis program, 5.Service processor, 6.Logging section, 7.Fault information analysis section,
8...Timer, 9...Counter, 10...Fault information, 11...Micro diagnosis program selection information.

Claims (1)

[Scope of Claims] A control storage (3) that stores a microprogram (3) that executes normal processing and a plurality of microdiagnosis programs (4A, 4B...4N) that executes fault diagnosis processing only when a hardware failure occurs. 2); a logging unit (6) that logs failure information (10) generated in the computer body when a hardware failure occurs; and a logging unit (6) that analyzes the logging information from the logging unit and performs the microdiagnosis described above. A failure information analysis unit (7) sends micro-diagnosis program selection information (11) to the computer main body for starting the most suitable diagnosis program from among the programs; A service processor (5) having a timer (8) for sending data to the computer main body, and a counter (9) for executing fault diagnosis by the optimal microdiagnosis program a predetermined number of times, 1. A fault diagnosis method for a computer system, characterized in that fault diagnosis based on an optimal micro-diagnosis program is automatically repeated a predetermined number of times at fixed time intervals only at certain times.