JPH064334A - Fault diagnostic system for logic device - Google Patents

Abstract

PURPOSE:To acquire suspected fault part name information by performing log analysis again when a service processor is restarted even in the case of stopping service before the log analysis is completed by holding information showing which log is completely analyzed concerning the log analysis of a logic device. CONSTITUTION:A log file 47 stores the information of collected log numbers, the information of analyzed log numbers, log information and the suspected fault part name information. When the service processor is stopped and restarted during log information #6 analysis, a log analyzing means 43 compares the analyzed log numbers with the collected log numbers at the time of restarting the service processor and when they are not coincident, a log analysis restarting means 46 is activated. The log analysis restarting means 46 activates the log analyzing means 43 to the log information #6. The log analyzing means 43 analyzes the log information #6 and writes the suspected fault part name information in the suspected fault part name information #6 area of the log file 47.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a failure diagnosis system for a logic device, and more particularly to a failure diagnosis system for a logic device in which a service processor performs a failure diagnosis for a logic device which constitutes an information processing system.

[0002]

2. Description of the Related Art Conventionally, a built-in diagnostic method (Built-in Diagnostics;
Hereinafter, the BID method) is known.

In this BID system, various error detection circuits by hardware are provided at important points in a logical unit such as an arithmetic processing unit, an input / output control unit, a main storage unit, etc. which constitute an information processing system, and an error detection circuit is provided. Error indicator flip-flop (hereinafter abbreviated as EIF) provided for error detection when the detection circuit detects an error
Is collected as an error log by the log analysis means in the service processor in the information processing system via the diagnostic unit and written in the magnetic disk device.

After that, the log collecting means activates the log analyzing means in the service processor for the collected error log. The log analysis means reads the error log corresponding to the specified log number from the magnetic disk device and performs a failure diagnosis. This failure diagnosis analysis is carried out by using the collected error log as a failure sign, and the suspected failure component name information obtained by this failure diagnosis analysis is written in the error log file of the magnetic disk device or the like. The suspected failure part name information written in the error log file is edited together with the error log and edited together with the error log as necessary, and is output.

[0005]

However, in the above-described conventional failure diagnosis method for a logical device, when the service processor is restarted after the error log is collected and before the log analysis means is terminated. Since there is no timing for activating the log analysis means, there is a problem that the error log is not analyzed and the log analysis result cannot be obtained.

[0006]

According to the method of the present invention, a service processor having a maintenance / diagnosis function, an error detection circuit provided in a key part of a logic device, and the error detection circuit detect an error of the logic device. In a failure diagnosis method of a logic device in an information processing device, which includes an error indicator flip corresponding to each error detection circuit set when Log collection means for collecting error logs in the error log file in the service processor in order of failure occurrence time via a diagnostic path, and collected log number information indicating up to which error log has been collected by the log collection means And update the collected number information when log collection is completed for the error log Collected log number information updating means and failure diagnosis analysis using the error log collected by the log collecting means as a failure symptom to point out suspicious failed part name information, and put the suspicious failed part name information in the error log file. Log analysis means for recording, analyzed log number information indicating up to which error log has been analyzed by the log analysis means, and updated analyzed number information when log analysis is completed for the error log The analyzed log number information updating means, the collected log number information when the service processor is started up, and the analyzed log number information are compared, and if they do not match, the log analysis means is started again. A logical unit failure diagnosis method characterized by comprising a restarting means.

[0007]

EXAMPLES Next, examples of the present invention will be described.

FIG. 2 is a block diagram showing an example of an information system to which the present invention is applied.
An arithmetic processing unit (CPU) 11 and an input control unit (IO
P) 12, main memory unit (MMU) 13, and diagnostic path 1
A diagnostic unit (DGP) 1 connected to the arithmetic processing unit 11, the input / output control unit 12, and the main storage unit 13 by means of 9.
4, a service processor (SVP) 15, a magnetic disk device 16, a console 17 and a printer 18.

The magnetic disk device 16 has a log number information updating means 41 and a log analyzing means 4 as shown in FIG.
3, collected log number information updating means 42, analyzed log number updating means 44, suspicious failed component name edit output means 45,
A log analysis restart means 46 and a log file 47 are stored. Further, in the log file 47, as shown in FIG. 4, collected log number information 51, analyzed log number information 52, and log information # 1 to #n (531 to 53n: n).
Corresponds to the log number) and suspected faulty part name information # 1 to #
n (541 to n) are stored.

FIG. 1 is a flow chart of an embodiment of the present invention. First, a procedure for performing a normal log analysis will be described. It is assumed that four error logs have already been collected in the log file 47, and the collected log number information 51 and the analyzed log number information 52 are both "4". If any failure occurs in the information processing shown in FIG. 2, the service processor 15 is notified via the diagnostic unit 14 (S6). Next, the service processor 15 activates the log collecting means 41 (S7).

The log collecting means 41 reads the failure information from the failure generating device via the diagnostic path 19 and stores it in the log information # 5 area in the magnetic disk device 16 connected to the service processor 15. Next, the log collecting means 41
Activates the collected log number information updating means 42. The collected log number information updating means 42 updates the collected log number information in the log file to "5" (S8). Next, the log collection unit 41 activates the log analysis unit 42. The log analysis means 42 starts analysis on the log information # 5 (S9) and writes it in the suspected faulty part name information # 6 area.

The log analysis means 43 then activates the analyzed log number information means 44. The analyzed log number information updating means 44 updates the analyzed log number information in the log file 47 to "5" (S10).

After that, the suspected failed component name display means 45 is activated as necessary, and the suspected failed component name information is output to the printer 18 or the console 17.

If the service processor 15 is restarted (S1), the analyzed log number information 52 and the collected log number information match each other, and therefore the log is kept until a new failure occurs. The analysis means 43 is not activated.

Now, assuming that the event shown in FIG. 5 has occurred, the operation of this embodiment will be described. First, a failure occurs (e1), log collection is completed (e2), log analysis is started (e3), the service processor 15 is stopped (e4), and then the service processor 15 is restarted (e5). ), The log analysis means 43 is restarted (e
6), the case where the log analysis is completed (e7) will be described as an example.

The error log in the log file 47 is already 5
It is assumed that the items have been collected and the collected log number information and the analyzed log number information are both “5”. If any failure occurs in the information processing apparatus shown in FIG. 2, the service processor 15 is notified via the diagnostic unit 14 (S
6). Next, the service processor 15 uses the log collecting means 4
1 is activated (S7).

The log collecting means 41 reads the fault from the fault occurrence device via the diagnostic path 19 and stores it in the log information # 6 area in the magnetic disk device 16 connected to the service processor 15. Next, the log collecting unit 41 activates the collected log number updating unit 42. The collected log number information updating means 42 updates the collected log number information in the log file 47 to "6" (S8). Next, the log collection unit 41 activates the log analysis unit 42. The log analysis unit 42 starts analysis on the log information # 6 (S
9). It is assumed that the service processor 15 stops during the log analysis as indicated by e4 in FIG.

After that, the service processor is restarted (S1). The service processor 15 compares the analyzed log number 51 and the collected log number 52 at the time of restart (S2). If they do not match as in the case of FIG. 5, the log analysis restart means 46 is started. The log analysis restart means 46 starts the log analysis means 43 for the log information # 6. The log analysis unit 43 analyzes the log information # 6 and writes the suspected faulty part name information in the suspected faulty part name information # 6 area in the log file 47 (S4).

The log analyzing means 43 then activates the analyzed log number information updating means 44. The analyzed log number information updating means 44 updates the analyzed log number information in the log file 47 to "6" (S5).

Thereafter, if necessary, the suspicious failed component name means 45 is activated and the suspicious failed component name information is sent to the printer 18.
Alternatively, it is output to the console 17. After that, when the service processor 15 is restarted, after that,
If it is assumed that the service processor 15 has been restarted (S1), the analyzed log number information 52 and the collected log number number information match each other, so the log analysis means 42 continues until the next new failure occurs. Will not start.

[0021]

As described above, according to the present invention, regarding the log analysis of the logical device, by having information on which error log has already been log analyzed, even if the service is stopped before the end of the log analysis, the service is provided. If the processor is restarted, there is an advantage that the log analysis is performed again and the suspicious failed component name information is obtained.

[Brief description of drawings]

FIG. 1 is a flowchart of an embodiment of the present invention.

FIG. 2 is a block diagram showing an example of an information processing system to which an embodiment of the present invention is applied.

FIG. 3 is a diagram showing stored contents of a magnetic disk device according to the present invention.

FIG. 4 is an explanatory diagram of contents of a log file according to the present invention.

FIG. 5 is a diagram showing an example of a transition diagram of collected log number information and analyzed log number information for explaining the operation of the present invention.

[Explanation of symbols]

 11 Exercise Processing Unit (CPU) 12 Input / Output Control Unit (IOP) 13 Main Memory Unit (MMU) 14 Diagnostic Unit (DGP) 15 Service Processor (SVP) 16 Magnetic Disk Unit 17 Console 18 Printer 19 Diagnostic Path

Claims (1)

[Claims] 1. A service processor having a maintenance / diagnosis function, and an error detection circuit provided in a key part of a logic device.
In a failure diagnosis method of a logic device in an information processing device, which comprises an error indicator flip corresponding to each error detection circuit set when the error detection circuit detects an error of the logic device, a failure occurs in the logic device. At this time, an error log containing the content held by the error indicator flip-flop, a log collecting unit that collects the error log file in the service processor in the order of failure occurrence time via a diagnostic path, and which error log is generated by the log collecting unit. Collected log number information indicating whether up to the above has been collected, collected log number information updating means for updating the collected number information when the log collection for the error log is completed, and the collected log number information by the log collecting means. Information on the name of the suspected faulty component is obtained by performing fault diagnosis analysis using the error log as a fault symptom. And a log analysis means for recording the suspected faulty part name information in the error log file, and analyzed log number information indicating up to which error log has been analyzed by the log analysis means, and the error log On the other hand, the analyzed log number information updating means for updating the analyzed number information when the log analysis is completed, and the collected log number information when the service processor is started up are compared with the analyzed log number information. And a log analysis restarting means for restarting the log analyzing means again when they do not match.