JP2702432B2 - Failure diagnosis method - Google Patents

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, and more particularly, to a system for specifying a failure package component by using a failure dictionary in an information processing system including a plurality of processing devices mounted with a large number of package components. The present invention relates to a built-in failure diagnosis method.

[0002]

2. Description of the Related Art In a conventional failure diagnosis system of this kind, two dictionary files, a failure diagnosis dictionary file prepared in advance and a mounting position / mounting parts dictionary file, are used.
He points out failed parts (package parts) at the time of hardware failure. The failure diagnosis dictionary file is a dictionary that is searched using a plurality of failure information flags, and has a configuration such as a failure information / failure point list that can indicate the mounting position of the failed hardware by a combination of the failure information flags. The mounting location / mounting component dictionary file is a file with a configuration like a list that shows which components are mounted at each mounting location. The mounting location output as a search result of the failure diagnosis dictionary file is searched using the search key. This is a dictionary file for indicating the type (article name) of a failed component, that is, a component that needs to be replaced.

In the early information processing systems, the components mounted in one mounting position were not changed by the system model or type.
A failure diagnosis method in which a failure diagnosis dictionary file directly points to a failed component without using a mounted component dictionary file has been adopted. In recent information processing systems, many models / types are set in one system according to differences in input / output performance, arithmetic performance, optional functions, and the like. Some of the components constituting the plurality of models / types are shared among the models / types and some are unique. Different components are mounted at the same mounting position of the same processing apparatus due to the difference of the models / types. Cases are increasing. Therefore, when a system failure diagnosis is performed, a failure diagnosis dictionary file must be created for each model / type in a system in which a failure component is directly pointed out using only the failure diagnosis dictionary file. However, since the size of the failure diagnosis dictionary file is generally large, the faulty component indication part is configured as a mounting position / mounting part dictionary file in a separate file, and only this mounting position / mounting part dictionary file is created for each model / type. However, the number of replacement methods has increased. In particular, when a system built-in failure diagnosis system is employed, this system is generally used.

However, in recent information processing systems,
With the advancement of multiprocessors and high parallelism, the number of models / types has become extremely large with the advancement and large-scale of the system scale. An increase in the number of component dictionary files is becoming a problem. Among them, as a countermeasure against the former, for example, Japanese Unexamined Patent Application Publication No. 3-278239 and Japanese Unexamined Patent Application Publication No. 2-181837 disclose proposals that can be applied when a plurality of devices having the same configuration such as a storage device are connected. However, no effective countermeasures have been shown to address the latter problem.

[0005]

As described above, when a failure diagnosis is performed for an apparatus having different mounting components due to a difference in model / type or the like, in the conventional failure diagnosis method, the mounting position / mounting component dictionary file contains the model / type. There are drawbacks that need to be made as many as. In particular, the number of models / types in a highly parallel system, which has recently become mainstream, is extremely large. Therefore, the amount of maintenance diagnosis design increases due to an increase in the number of mounting locations and mounting component dictionary files, and in consideration of future system expansion. Therefore, it is difficult to store the mounting position / mounting component dictionary file of each model / type in a disk device in advance.

SUMMARY OF THE INVENTION An object of the present invention is to provide a failure diagnosis method which can eliminate the above-mentioned disadvantages, and is particularly effective when applied to an information processing system in which package components mounted at the same position differ depending on models / types. It is in.

[0007]

According to a first aspect of the present invention, there is provided a failure diagnosis system for performing a failure diagnosis by connecting an arbitrary number of processing units having a plurality of replaceable package parts mounted thereto and the processing units via a diagnosis bus. An information processing system comprising: a diagnostic processor; and a service processor that searches a failure dictionary based on the diagnostic information from the diagnostic processor, and outputs mounting position information and component identification information of a package component that is suspected of failure and is to be replaced. In the failure diagnosis method, the failure dictionary is configured with only a failure diagnosis dictionary file that searches with diagnostic information from the diagnostic processor and outputs suspected mounting position information on which the suspected package component is mounted, and is mounted on the processing device. In addition to a failure detection circuit that detects the failure of the package component, Fault reporting means comprising identification information holding means for storing and holding different information, wherein the processing device collects the output of the failure detection circuit of each package component mounted on the processing device and reports the output as diagnostic information to the diagnostic processor; And selecting means for controlling reading of the component identification information from the identification information holding means, wherein the diagnosis processor notifies the service processor of the diagnosis information from the failure reporting means and responds according to an instruction from the service processor. And a function of instructing the selection unit of the processing device to perform the readout position.

According to a second aspect of the present invention, in the fault diagnosing method according to the first aspect, the service processor comprises:
After instructing the diagnostic processor to execute a failure diagnosis, instructing the diagnostic processor to read out component identification information based on suspected mounting position information retrieved based on the diagnostic information from the diagnostic processor. .

According to a third aspect of the present invention, in the failure diagnosis system according to the first aspect, the service processor comprises:
When the information processing system is started, the component identification information of each package component is sequentially read out to create a correspondence list of mounting position information and component identification information, and at the time of failure diagnosis, the created correspondence list is searched and necessary components are searched. It is characterized by reading the identification information.

[0010]

Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of one embodiment of the present invention.

As shown in FIG. 1, the failure diagnosis system of this embodiment includes four processing units 1, 2, 3 and 4 on which a plurality of replaceable package parts are mounted, and these units are connected to each other via a diagnostic bus. A diagnostic processor 5 connected to perform failure diagnosis, and a disk device 9 based on diagnostic information from the diagnostic processor 5
It comprises a service processor 6 for searching the failure diagnosis dictionary file 91 and outputting mounting position information of the suspected component, a maintenance screen device 7, and a printer device 8.

The processing units 1, 2, 3 and 4 are main processors of a multiprocessor information processing system having a plurality of models, each of which performs data processing, arithmetic processing, etc. A plurality of possible package components are mounted. This information processing system has four types of models as shown in FIG. 2, and the configuration of FIG. 1 having four processing devices is the largest model.

FIG. 2 is an explanatory diagram showing the configuration of a processing device corresponding to each model of the information processing system shown in FIG. In the figure, the processing device shown by a solid line represents an existing device, and the processing device shown by a dotted line represents a device that can be expanded in the future. That is, in the case of the minimum model M1, only the processing device 1 indicated by the solid line is configured, and the processing devices 2, 3 indicated by the dotted line
And 4 do not exist. In the case of the maximum model M4,
This indicates that all of the processing devices 1, 2, 3, and 4 indicated by solid lines exist. Similarly, intermediate models M2 and M3 located between the minimum model M1 and the maximum model M4 are shown to be composed of two or three processing devices indicated by solid lines, respectively. Therefore, FIG. 1 is a system of the maximum model M4.

Next, mounting of package components in each of the processing devices 1, 2, 3, and 4 will be described. Although a plurality of package components are mounted on the processing devices 1, 2, 3, and 4, respectively, only the processing device 1 is shown in detail in FIG. Inside the processing device 1, packages 10, 20, 30, and 40 are mounted as replaceable package components. Detection circuits 11, 21, 31, and 41 for detecting occurrence of an abnormality in the package are provided in each package.
And component information 12, 22, which can identify the package name
32 and 42 are mounted. The package 40 further includes detection circuits 11, 21, 31, and 41 of each package.
And a report circuit 43 for generating a diagnostic information signal of the processing apparatus 1 by inputting an output signal from the processor 1 and outputting the diagnostic information signal to the diagnostic processor 5, and component information 12, 22, A selection circuit 44 for selecting a corresponding one from 32 and 42 and outputting it to the diagnostic processor 5 is provided as a common circuit.

Here, the component information 12, 22, 32 and 42
Will be described in more detail. Parts information 12, 22, 3
Numerals 2 and 42 are, for example, 4-bit component identification information, and are set so that the component information displayed in hexadecimal and the specific package name correspond to each other, as shown in FIG. In the device configuration of each model shown in FIG. 2, the package components mounted on the processing device 1 included in all of the minimum model M1 to the maximum model M4 differ depending on the model.
The package names of the package components mounted at the mounting positions corresponding to 0, 30, and 40 are as shown in FIG. 4 for each model. That is, the maximum model M shown in FIG.
4, the package 10 of the processing device 1
The package names of 20, 30, and 40 are respectively AA
2, AA3, AA4, and BB3, and corresponding component information 12, 22, 32, and 42 in hexadecimal notation (1) _H , (2) _H , and (3) on each package component, respectively.
_H and (A) Binary information corresponding to _H is stored.
On the other hand, in the case of the minimum model M1, the package names are AA1, AA1, AA1 and BB1, and the hexadecimal numbers (0) _H , (0) _H , (0 ) _H and (8) 2 for _H
Stores value information. As described above, the component information is described as 4-bit information. However, when the number of types of package components to be identified is more than 16, it can be dealt with by increasing the number of bits. The component information needs to be stored as non-volatile information on each package component, but can be easily realized by using a small number of gate circuit elements.

The diagnostic processor 5 includes processing units 1, 2, 3,
And 4, which is activated by an instruction from the service processor 6 when a failure occurs, and executes a built-in diagnostic program for the processing devices 1, 2, 3, and 4. The diagnostic information signals obtained from the processing units 1, 2, 3, and 4 corresponding to each processing stage of the diagnostic program are transferred to the service processor 6 for failure diagnosis. The diagnostic program executed by the diagnostic processor 5 differs depending on the processing device to be diagnosed. In general, the same program is used for the same type of processing device (for example, the processing device 1 of each model shown in FIG. 2) regardless of the model / type. Can be applied. With respect to the processing device that is additionally provided depending on the model / type (for example, the processing device 2 with respect to the processing device 1 in FIG. 2), the number of components of the package components is not always the same, and may be different programs.

The service processor 6 comprises processing units 1 and
A processor for controlling fault diagnosis of 2, 3 and 4;
When a diagnostic information signal is received from the diagnostic processor 5, the failure diagnostic dictionary file 91 stored in the disk device 9 is searched based on the content of the diagnostic information signal, and the location of the suspected package component (suspect mounting position information). Recognize. Next, the service processor 6 instructs the diagnostic processor 5 to read out the component information of the package component mounted at the location. The component identification information of the suspected fault component thus obtained is output to the maintenance screen device 7 and the printer device 8 together with the mounting position information according to the instruction of the service processor 6.

As described above, each package component mounted on the processing device stores component information for identifying the product name. If the mounting position information of the suspected component is found as a result of the failure diagnosis, it is necessary to replace the component. The name of the package component to be prepared can be directly obtained. Therefore, it is only necessary to create a failure diagnosis dictionary file for determining the mounting position based on the diagnostic information. Even when the package component to be mounted is changed depending on the model / type, there are several types of mounting positions and mounting components for failure diagnosis. There is no need to create a dictionary file, which reduces the number of design steps and the storage area.

In the above description, the part information is read out from the package part at the corresponding position after obtaining the position information of the suspected part by executing the diagnostic program. However, when the information processing system is started, the part information is read from all the package parts. The configuration may be such that the correspondence list of the mounting position and the package name is created and stored sequentially, and the package name is read from the correspondence list when the diagnostic program is executed.

In the above-described embodiment, the report circuit for generating the diagnostic information signal and outputting the signal to the diagnostic processor and the selecting circuit for selecting the component information and outputting it to the diagnostic processor are provided in one of the package components. However, it may be provided not on the package component to be replaced but on the main body side of the processing apparatus.

[0022]

As described above, the failure diagnosis system of the present invention is configured to read component information that can specify a package name from actually mounted package components and output the information as suspected failure component information. Because
Even in the case of a system in which mounting parts (packages) are changed for a plurality of models / types, there is no need to create a mounting position / mounting parts dictionary file for each model / type, and one failure diagnosis dictionary file common to the models / types There is a great effect that an effective failure diagnosis can be performed only by using the above.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a device configuration for each model of the information processing system.

FIG. 3 is an explanatory diagram showing correspondence between component information and package names.

FIG. 4 is an explanatory diagram showing a package component configuration for each model of the processing apparatus.

[Explanation of symbols]

 1-4 processing device 5 diagnostic processor 6 service processor 7 maintenance screen device 8 printer device 9 disk device 10, 20, 30, 40 package 11, 21, 31, 41 detection circuit 12, 22, 32, 42 part information 43 report circuit 44 Selection circuit 91 Failure diagnosis dictionary file

Claims (3)

(57) [Claims] An arbitrary number of processing devices mounted with a plurality of replaceable package parts, a diagnostic processor connected to the processing device via a diagnostic bus and performing a failure diagnosis, and based on diagnostic information from the diagnostic processor. A service processor that outputs a mounting location information and component identification information of a package component to be replaced with a suspected failure due to a failure dictionary. It consists of only a failure diagnosis dictionary file which searches with the diagnostic information from and outputs the suspected mounting position where the suspected package component is mounted, and each package component mounted on the processing device detects a failure of the package component. In addition to the failure detection circuit, identification information holding means for storing and holding component identification information of the package component is provided. A failure report unit that collects the output of the failure detection circuit of each package component mounted on the processing device and reports the result as diagnostic information to the diagnostic processor; and component identification information from the identification information holding unit. Selecting means for controlling the reading of
The diagnostic processor has a function of notifying the service processor of diagnostic information from the failure reporting unit and instructing the selecting unit of the corresponding processing device to read a position according to an instruction from the service processor. Failure diagnosis method. 2. The system according to claim 2, wherein said service processor instructs said diagnostic processor to execute a failure diagnosis, and said component identification information to said diagnostic processor based on suspected mounting position information retrieved based on diagnostic information from said diagnostic processor. 2. The failure diagnosis method according to claim 1, wherein a read instruction is issued. 3. The service processor sequentially reads out component identification information of each package component at the time of starting up the information processing system, creates a correspondence list of mounting position information and component identification information, and creates the corresponding list at the time of executing a failure diagnosis. 2. The failure diagnosis method according to claim 1, wherein the correspondence list is searched to read out necessary component identification information.