JPH025147A - System for diagnosing logical device - Google Patents

Abstract

PURPOSE:To simplify the formation of a fault diagnostic dictionary by newly calculating and outputting suspected fault probability in which a common suspected fault part is failed when the same suspected fault parts exist. CONSTITUTION:The fault diagnosis dictionary 603 stores suspected fault parts information including the identification (ID) information and the suspected fault probability in which the suspected fault parts are failed in accordance with respective error indicator flip flops (EIFs). At the time of diagnosing a fault turning plural EIFs to a set state, a service processor 5 newly finds out a suspected fault probability in which suspected fault parts appearing in individual suspected parts information in common are failed based on suspected parts information stored in the dictionary 603 correspondingly to the EIFs turned to the set state. The found suspected fault probability is outputted from an output device 7 together with the ID information of the suspected fault part. Consequently, the formation of the fault diagnostic dictionary can be simplified.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a method for diagnosing a logic device, and in particular, the present invention relates to a method for diagnosing a logic device, and in particular, diagnosis of a logic device based on the state of an error indicator flip-flop included in the logic device and the contents of a fault diagnosis dictionary. This invention relates to a diagnostic method for logic devices that identifies suspected malfunctioning components.

[Prior Art] When a failure occurs in a logical device such as an arithmetic processing unit, main storage device, or input/output control device, the logical device is configured to facilitate investigation of the cause and enable prompt maintenance. independently replaceable parts (packages)
A necessary and sufficient number of check circuits and an error indicator 71J block flop (hereinafter referred to as EIF) that is set according to the error detection of this check circuit are provided in the internal circuit, and the clock of the logic device is stopped when a failure occurs. death,
The service processor checks which EIF has been set, searches the fault diagnosis dictionary using the set EIF, finds the part that contains the faulty part, outputs it from the output device, and then the maintenance person replaces this part. Conventionally, maintenance and repair of faulty parts have been carried out by

Conventionally, the above-mentioned fault diagnosis dictionary includes the name of the suspected faulty part, the mounting position of this part, and the level of probability that the part is faulty (for example, high, medium).

The suspected parts information consisting of 3 levels of low
In addition, the name of the suspected faulty part, the mounting position of this part, and the level of probability that the part is faulty (for example, high, medium,
It has suspect part information consisting of 3 levels (low and low),
If the service processor detects that multiple PIFs are in the server state when a failure occurs, the service processor controls the server.

The suspect component information stored in the above-mentioned failure diagnosis dictionary is searched for the combination of EiFs that have failed, and the suspect component name and its mounting position indicated by this suspect component information are searched according to their probability level. , parts with high and low probability of suspected failure were divided into levels and outputted from the output device.

[Problem to be solved by the invention]

The conventional logic device diagnosis method described above can automatically identify suspected failed parts that match the failure when multiple EIFs are set, which reduces the burden on maintenance staff. do. In other words, when multiple EIFs are in the set state, it is not possible for a maintenance person to simply output information on each suspected component corresponding to the EIF in the sent state to the output device. Although the maintenance person must determine by himself which parts are likely to be faulty, by configuring as described above, the maintenance person can immediately recognize the parts that are likely to be faulty. However, since it is necessary to store suspect component information in advance in the fault diagnosis dictionary in accordance with the EIF combination, the creator of the fault diagnosis dictionary must investigate and study the EIF combinations and their suspect component information in detail. Therefore, it takes a lot of time to create a fault diagnosis dictionary.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to identify suspected failure parts that match the combination of EIFs when a failure occurs where a plurality of EIFs are set, without having to store suspect part information corresponding to a combination of EIFs in a failure diagnosis dictionary. The object of the present invention is to provide a diagnostic method for logical devices that can make indications.

(Means for Solving the Problems) In order to achieve the above object, the present invention includes a logic device including an EIF that is set in response to error detection in a check circuit when a failure occurs, and a logic device that is connected to the logic device by a diagnostic path. A diagnostic method for a logic device, in which the service processor points out a suspected faulty component based on EIF status information collected via the diagnostic bus when a failure occurs, The service processor has a fault diagnosis dictionary that stores suspected part information including identification information of a faulty part and a suspected failure probability that the suspected faulty part has failed, and the service processor is configured to detect a fault in which a plurality of EIFs are set. At the time of diagnosis, E in each set state
Based on the suspect component information stored in the failure diagnosis dictionary in correspondence with the IF, newly determine the probability of a suspected failure of a suspected failure component that commonly appears in separate suspect component information;
The system is configured to output the obtained suspected failure probability together with identification information of the suspected failure component from an output device.

[Effect]

In the logic device diagnosis method of the present invention, the fault diagnosis dictionary stores suspected part information including identification information of a suspected faulty part and a suspected failure probability that the suspected faulty part has failed, corresponding to each ETF. , when diagnosing a failure in which a plurality of ErFs are in the set state, the service processor, based on the suspected part information stored in the fault diagnosis dictionary corresponding to each EIF in the set state, selects different suspected parts information. The suspected failure probability of the suspected failure parts that commonly appear in the failure parts is newly determined, and the calculated suspected failure probability is outputted from the output device together with the identification information of the suspected failure parts.

[Example] Next, an example of the present invention will be described in detail with reference to the drawings.

Referring to FIG. 1, an example of an information processing system to which the logical device diagnosis method of the present invention is applied is a system control device 1.
, a main storage device 2 , an arithmetic processing device 3 , an input/output control device 4 , a service processor 5 connected to these via a diagnostic path 100 , and a magnetic disk device 6 connected to this service processor 5 . It is composed of a printer 7 and an operation console 8. The magnetic disk device 6 includes a fault diagnosis dictionary 603 and an E
(2) An F information storage area 601 and a diagnosis result storage area 602 for holding the results of diagnosis based on the collected EIF information and the fault diagnosis dictionary 603 are provided. In addition, when a failure occurs in a logical device such as the arithmetic processing unit 3, the service processor 5 collects EIF information from the failed logical device via the diagnostic path 100 and collects the EIF information from the magnetic disk. EIF information storage area 601 of device 6
Diagnosis is performed using the collection program 501 stored in the EIF information storage area 601 and the EIF information held in the EIF information storage area 601 and the failure diagnosis dictionary 603, and the results are stored in the diagnosis result storage area 602.
The diagnostic program 502 stored in the console 8, the diagnostic results stored in the diagnostic result storage area 602 of the magnetic disk device 6 in response to operations on the console 8, and the EIF information stored in the ErF information storage area 601 as needed. A display program 503 for editing and outputting to a printer or displaying on the display device of the console 8 is built-in.

Referring to FIG. 2, the fault diagnosis dictionary 603 includes each EIF
It has a logical structure that stores one or more suspect component information corresponding to the suspected failure component, and each suspect component information includes the name of the suspected failure component, the mounting position of this suspected failure component, and the location of the failure of this suspected failure component. It consists of the probability of a suspected failure (expressed in %, for example).

Note that identification information of the suspected failed component is composed of the suspected failed component name and the mounting position. Generally, in one information processing system or one logical device, many parts (pancages) with the same name are used, and individual parts cannot be identified by the part name alone, so in this example, the mounting position is also added. This is used as part identification information. Therefore, if the component name is unique or if the component can be identified only by the mounting position, the identification information of the suspected faulty component may be composed of only the suspected faulty component name or only the mounting position.

Each suspected part information illustrated in FIG. 2 corresponds to a hardware configuration, for example, as shown in FIG. 3, which exists inside the arithmetic processing unit 3 constituting one of the logical devices. That is,
Component PKGI at mounting position 11ffiL1 including register 10.11 and component P at mounting position L2 including register 20
KG2, a component PKG3 at a mounting position L3 including a register 30, a parity check circuit 31, and an EIF32,
Registers 40-42. Selector 43. Implemented IL including parity check circuit 4445 and EIF46.47
4 component PKG4 exists, and each element is connected as shown in the figure, the parity check circuit 4
4, when a parity error is detected and the EIF46 is set, the component PKC; 1. Since PKG4 is considered to be a suspected faulty component, two pieces of suspect component information including its suspected failure probability are stored in the fault diagnosis dictionary 603 in correspondence with E[F46 and are processed by the parity check circuit 31 as shown in FIG. When a parity error is detected and EIF32 is set, parts PKG2 and PKG3 are considered to be the suspected failure parts, so 2
As shown in FIG. 2, information on two suspected parts is stored in the failure diagnosis dictionary 603 in correspondence with <EIF32, and when a parity error is detected in the parity check circuit 45 and ErF47 is set, the parts PKCI, PKG3. P.K.
Since G4 is considered to be the suspected faulty part, information on the three suspected parts including its suspected failure probability is not shown in Figure 2.
It is stored in the fault diagnosis dictionary 603 corresponding to F47. The suspected failure probability in each suspected component information is determined based on the number of elements in each component that affects the parity check circuit, the designer's experience, and the like.

Next, the operation of the embodiment shown in FIG. 1 will be described below, taking as an example a case where a failure occurs in the arithmetic processing unit 3.

Referring to FIG. 1, when a failure occurs in the processing unit 3, the service processor 5 collects EIF information indicating the status of each EIF from the processing unit 3 via the diagnostic path 100 by executing the collection program 501. collect and
It is stored in the EIF information storage area 601 of the magnetic disk device 6. Next, the service processor 5 uses the diagnostic program 502
Diagnosis is performed by executing

Referring to FIG. 4, the diagnosis by the diagnostic program 502 first obtains one EIF indicating the cent state from the ETF information storage area 601 (S 1 ), and uses this obtained EIF of the set state as an index to perform the diagnosis in FIG. The indicated failure diagnosis dictionary 603 is searched, the corresponding suspect part information is acquired, and this is stored in the diagnosis result storage area 602 (S3). ) Repeat for the EIF of state. Then, when the acquisition and storage of suspect parts information is completed for all EIFs in the set state (if YES at 32), the EIFs in the set state
It is determined whether or not a plurality of F exists (S4), and if there is not, the process is immediately terminated. Conversely,
If there are multiple EIFs in the set state, the diagnostic result storage area 602 is used for each EIF in the set state.
Analyze each suspect part information stored in the , and if there is a suspected faulty part that appears in common in separate suspect parts information, calculate the new suspected failure probability of that suspected faulty part,
The obtained suspected failure probability is stored in the diagnostic result storage area 602 together with the name of the suspected failure component and the mounting position (S5).

Then, the process ends.

Thereafter, when a request to display diagnostic results is input from the console 8, the service pro sensor 5 edits the diagnostic results stored in the diagnostic result storage area 602 of the magnetic disk device 6 by executing the display program 503. and output from the printer 7. Therefore, in this example, as a diagnosis result,
Suspected part information corresponding to the EIF that became 7-bit status,
If there are multiple EIFs in the set state and there is a common suspected faulty part, the suspected fault probability newly determined in step S5 of the diagnostic program 502, the name of the suspected faulty part, and the mounting position are sent from the printer 7. It will be output.

Next, the process of step S5 of the diagnostic program 502 will be explained in more detail.

For example, in the hardware configuration shown in Figure 3, EIF4
EIF 6 and EIF47 are both set.
If the information is collected, referring to Figure 2, EIF4
Corresponding to ErF47, two pieces of suspect component information are obtained indicating that component PKG4 at mounting position L4 has a suspected failure probability of 60% and component PKGI at mounting position L1 has a suspected failure probability of 40%. Component PKG at mounting position L4
4 is a component PKG with a suspected failure probability of 60% and an execution position zLt.
I is a component PKG with a suspected failure probability of 25% and a mounting position L3.
Information on three suspect parts is obtained indicating that PKG4.3 has a suspected failure probability of 15%. PKC;1 is commonly included.

Therefore, the suspected failure probability of component PKG4 and the suspected failure probability of component PKG1 are determined by the following equations.

・Suspected failure probability of component PKG4 - (Part PKG in suspect component information compatible with EIF46
4 suspected failure probability 60% + Part PKG4 in suspect part information compatible with ETF41
60% probability of suspected failure) ÷ (P in suspected parts information compatible with E I F 4.6
100% sum of suspected failure probabilities of all suspected failure parts commonly included in IF46.47 + EIF46 in suspect part information compatible with EIF47,
85% sum of suspected failure probabilities of all suspected failure parts that are commonly included in ErF47)
40% probability of suspected failure of I
25% probability of suspected failure) + (EIF46 in suspected parts information compatible with EIF46)
．． 100% sum of suspected failure probabilities of all suspected failure parts commonly included in EIF47 + EIF46 in suspected parts information compatible with EIF47,
The sum of the suspected failure probabilities of all suspected failure parts commonly included in 47 (85%)
The mounting position Ll and the suspected failure probability of 35% obtained above are stored in the diagnostic result storage area 602 for later output. As a result, the maintenance person determines that if there is enough maintenance time, parts PKG4. It can be seen that it is better to replace the parts one by one in the order of parts PKCI and perform the test, and that it is better to replace parts PKG4 and parts PKCI at the same time if the maintenance time is short.

Furthermore, in the hardware configuration shown in Fig. 3, E[F32
and EIF47 are both set.
When the F information is collected, the component PKG3 at the mounting position L3 has a suspected failure probability of 10 using the same calculation method as described above.
0%, and this will be indicated in the diagnostic results. Therefore, the maintenance person can determine that only component PKG3 needs to be replaced.

Although the embodiments of the present invention have been described above, the present invention is not limited to the above embodiments, and various other additions and changes are possible. For example, when EIF information is collected in which multiple Snake IFs are set at the same time, in addition to the newly determined suspected failure probability and the identification information of the suspected failed component in step S5 of FIG. Although the suspect part information is also output as the diagnosis result, the latter output of the suspect part information may be omitted.

〔Effect of the invention〕

As explained above, in the logic device diagnostic method of the present invention, when a plurality of EIFs are in the set state,
The information on each EIF-compatible suspected component is compared, and if the same suspected failed component exists, the common suspected failed component is determined to have failed based on the suspected failure probability indicated in the individual EIF-compatible suspect component information. Since the probability of suspected failure is newly calculated and outputted, there is no need for the failure diagnosis dictionary to include information on suspected parts corresponding to the combination of EIFs, and the creation of the failure diagnosis dictionary becomes simple. In addition, since the probability of suspected failure of the components is shown not in high, medium, or low levels, but in more detailed values, what kind of replacement procedure should be adopted depending on the maintenance situation? It is also possible to make detailed judgments.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an example of an information processing system to which the present invention is applied, FIG. 2 is a diagram showing an example of the configuration of a fault diagnosis dictionary, and FIG.
A block diagram showing an example of the hardware configuration for which the failure diagnosis dictionary was created, and FIG.
2 is a flowchart of a processing example of No. 02. In the figure: ...System control device 2...Main storage device...
Arithmetic processing unit 4...Input/output control 1 device...
Service processor 6...Magnetic disk device...
Printer 8...Operation console 00...Diagnosis path Ol...Collection program 02...Diagnosis program 03...Display program 01...EIF information storage area 02...Diagnosis result storage area 03...・Fault diagnosis dictionary

Claims (1)

[Scope of Claims] A logic device including an error indicator flip-flop that is set in response to error detection of a check circuit when a failure occurs, and a service processor connected to the logic device by a diagnostic path, In a logic device diagnosis method in which the service processor points out a suspected faulty component based on status information of an error indicator flip-flop collected via a diagnostic path, identification information of the suspected faulty component is provided corresponding to each error indicator flip-flop. The service processor has a fault diagnosis dictionary that stores suspect component information including a suspected failure probability that the suspected failed component has failed, and when diagnosing a failure in which a plurality of error indicator flip-flops are set, Based on the suspect component information stored in the fault diagnosis dictionary corresponding to each error indicator flip-flop in the set state, the probability of a suspected failure of a suspect component commonly appearing in separate suspect component information is determined. A method for diagnosing a logic device, characterized in that the newly determined probability of a suspected failure is output from an output device along with identification information of the suspected failure component.