JPS6195457A - Diagnosis processing method - Google Patents

Abstract

PURPOSE:To point out the position of a failure automatically without manual operation by dividing and defining an error indicator FF(EIF) into a source EIF and a transmitted EIF and analyzing error log information to decide the source EIF. CONSTITUTION:An information processing system consisting of an operation processor (EPU)1, a service processor (SVP)5, etc. includes an error detecting circuit, which has an error indicator FF(EIF). The EIF is divided into the source EIF set up at first at the generation of an error and the transmitted EIF to be reset when the error is transmitted and the transmitted error is detected and the transmitted EIF and these EIFs are defined. When an error is generated in the EPU1, the information is informed from the EPU1 to the SVP5, which records all EIFs included in the EPU1 and the contents of other necessary registers as error log. In case of error diagnosis, the error log information is analyzed and the source EIF is decided, so that the position of the fault can be pointed out without using manual operation.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a diagnostic processing method, and in particular to a diagnostic processing method that collects an error log when an error occurs in an information processing system and points out the failure location by analyzing the error log. Regarding.

[Conventional technology]

In general, Ic in each device that constitutes an information processing system is
Equipped with various hardware error detection circuits,
An error indicator flip-flop (hereinafter referred to as EIF) is provided corresponding to each error detection circuit, which is set by VCG when these error detection circuits detect an error.

When an error occurs, an error log containing information indicating the set/reset status of each EIF is collected by, for example, a service processor in the system and recorded on a disk or the like.

When an error occurs, the error log collected in this way is edited appropriately and displayed on the console display or printed out as necessary, and the information indicating the set/reset status of each EIF mentioned above is analyzed by the operator. The fault location is diagnosed.

[Problem that the invention seeks to solve]

Now, when an error occurs, if the first error that occurs is detected by an error detection circuit whose scope is to detect this error, and only the EIF corresponding to this detection circuit is set, the fault location can be diagnosed. easily done.

However, in general, once an error occurs, this error propagates one after another, and is also detected as an error by the propagating light error detection circuit, resulting in each of the above-mentioned E
Information indicating the set/reset status of the IF indicates that an EIF set due to the detection of a propagated error is also displayed as a set EIF, so by analyzing this information, it is possible to diagnose the failure location that caused the error. It is often very difficult to do so.

In contrast, in the past, operators performed diagnosis using their own methods depending on each case, which had the disadvantage of consuming a great deal of labor and time.

SUMMARY OF THE INVENTION An object of the present invention is to provide a diagnostic processing method that eliminates the above-mentioned conventional drawbacks and automatically points out a failure location using an organized standard processing method without relying on human intervention.

[Means to solve all problems]

The system of the present invention includes a plurality of hardware error detection circuits, an error indicator flip-flop (hereinafter abbreviated as EIF) corresponding to the error detection circuit that is set in response to error detection by each of the error detection circuits, and an error detection circuit. an error log collecting means for collecting an error log including information indicating the set/reset state of each EIF when an error occurs, and analyzing information indicating the set/reset state of each EIF in the collected error log. In the diagnostic processing method that points out the failure location, the EIF that is initially set when an error occurs is called the original EIF.
E Primitive EI
Prepare error propagation information for all the EIF'i that can become F as the upper EIF' for the EIF, and set in the information indicating the set/reset status of each EIF in the collected error log. The upper EIF for each EIF in the error propagation information
If all IFs have been reset in the error log, the EIF is determined to be the original EIF, and the failure location corresponding to the EIF is pointed out.

〔Example" Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a block (2) showing one embodiment of the present invention.

this! The example is an arithmetic processing unit (EPU) 1, a system control unit (SCU) 2. Input/output control unit (IOP) 3, main memory unit (MMU) 4, service processor (SVP)
5, a disk 6, an encoder 7, and a printer 8.

These EPUI, 5CU2. l0P3. Various devices such as the MMU4 include a hardware error detection circuit that detects all error occurrences, and each error detection circuit sets an error corresponding to each error detection circuit when this circuit detects an error. It has indicator flip-flop (hereinafter referred to as EIF)=i.

An example of the arrangement of these error detection circuits and EIFs in each installation is shown in FIG. 2(A) VC. In this example, input information 100 is subjected to certain processing by process P1, and its output is processed by process P2 and process P5.
is supplied as input information.

The outputs a and 7'0 of process P2 and process P5 are supplied as inputs to process P3, and the results of the processing are supplied to process P4 and process P6vc, respectively.

Each of the above processes is processed by a hardware maintenance/replacement unit (hereinafter referred to as FRU) K. For example, the process PI is FRU12, FRU12, etc.
Processed by N8 units called F-UIN,
Generally, process Pi is FRUil.

F 几Ui2. It is assumed that processing is performed by N units called FRUiN4 (each FRU is not shown).

Now, each process pih includes an error detection circuit Di,
A unit FRUil performs processing of process P1,
FRUi2. ,...9.・If a failure occurs in any of F-UiNi, it will be detected by the detection circuit D1, and the corresponding EIFi will be set.

For example, now, a failure has occurred in the unit )FU21 in the process P2, so the error detection circuit D2 detects all errors and corresponding EIF2'! Suppose that it is set to l-. The output of process P2 contains errors due to the above-mentioned failure and is fed to process P31C.

As a result, an error is detected in the error detection circuit D3 due to input information containing an error that is supplied even though there is no failure in each unit (H bar) of process P3, and EIF3 is set. There may be cases where this happens.

Similarly, processes P4 and P6, which have received input information containing an error from process P3, have their respective error detection circuits D4. L) EIF4.2- is detected at EIF4.6. A case may occur where EIF6 is set.

In this way, not only is EIF2 corresponding to P2 that caused the actual failure in the hardware set.

Due to error propagation, EIF3. EIF4 and EIF6
This means that there is also a possibility that it will be set.

Now, like EIF2 in the above example, EIF'i is defined as the primitive EIF that is first set due to a failure of hardware included in its own area of responsibility, and EIF3 in the above example.
EIF4. Let us define EIFi set by error propagation as EIF6 as the propagated EIF for this original EIFK. In this way, a certain EIF becomes a primitive EI
When F is reached, EIFs that can be propagated EIFs corresponding to this can always be added in consideration of error propagation as described above.

Figure 2 ■ shows each EIF in the example shown in Figure 2 (A)K.
becomes a primitive EIF, each primitive EIFV
FIG. 3 is a table showing each EIF that can be a propagated EIF' for c.

Using this table, it is easy to create a table that lists, for each EIF, all possible EIFs that can be a primitive EIFK such that the EIF gold can be propagated P.

For example, from the table in Figure 2 ■, EIF3 is the propagated EI
The primitive EIFs listed as F include EIFl,
EIF2. EIF5 exists, indicating that these are EIFs that can be primitive EIFKs, such that EIF3 can be a propagated EIF.

In this way, for each EIF, it is possible to create a table shown in Figure 2 (O) that lists the EIFs that can be a primitive EIF such that the EIF can be a propagated EIF. For a certain propagated EIF, the primitive E
An EIF that can become an IF is always a $6EIF with an upper order in terms of error propagation, and henceforth, for simplicity, this will be referred to as an upper EIF. For example, EIFI is EIF
It is a higher EIF than '3. Also, the table shown in Figure 2 ◎ will be called the error propagation information table.

In the diagnostic processing method of this embodiment, an error propagation information table as shown in FIG. 2 is prepared for each EIFVc to be subjected to diagnostic processing.

Roughly speaking, in the diagnostic processing method of this embodiment, an EIF-FRU correspondence table is prepared for each EIF, listing the maintenance/replacement units (FRUs) that process the process corresponding to the EIFVC.

Furthermore, as will be described later, an EIF set/reset table indicating the set/reset status of each EIF when an error occurs is edited and created, and processing is executed with reference to these three types of tables.

FIG. 3 shows an example of these three types of tables, and FIG. 3A) is a fiEIF set/reset table, FIG. It is a figure which shows an example of a correspondence table.

Now, the diagnostic processing of this embodiment is executed as follows.

For example, if an error occurs in the EPU 1, information on the occurrence of the error is reported from the EPUI to the 5VP5.

Upon receiving this notification, 5VPs reads all EIFs included in the EPUI by forming a series of shift paths with other necessary registers included in the EPUI and shifting out their contents, and detects this as an error. The information is collected as a log and recorded on the disk 6.

Next, 5vps edits the error log thus collected and creates a table showing the set/reset status of each EIF included in this error log (the EI shown in Figure 3 above).
F set/reset table) Create all.

Now, the 5VP5 diagnostic program refers to this EIF set/reset table, the error propagation information table and the EIF-FRU correspondence table prepared in advance, and follows the flowchart shown in FIG.
Perform diagnostic processing.

That is, when this diagnostic program starts (the fourth
Figure A), first take the first EIF as the processing target (Figure 4 B). Next, refer to the EIF set/reset table mentioned above to check whether this EIF has been set (Figure 4), and if it is not set, proceed to Figure 4 (K). . If it is set, check whether an upper EIF of this EIF exists by referring to the above-mentioned error propagation information table. Since this means that an EIF that does not exist has been set, this EIF is naturally determined to be the primitive EIF, and the next check is skipped and the process proceeds to the fourth diagram. If an upper EIF exists, the error propagation information table and the EIF set/reset table are referred to to check whether all upper EIFs for this EIF have been reset (see Figure 4). If the top E
If there is something set in IF, this E
It is determined that the IF is not a primitive EIF, and the process proceeds to Figure 4.

On the other hand, if all the upper EIFs have been reset, the VC determines that this is the primitive EIF and proceeds to the next fourth diagram.

In the fourth diagram, this EIF is set,
Moreover, as mentioned above, this EIF is determined to be the original EIF because the upper EIF does not exist, or even if it does exist, there is nothing set in the upper EIF, and therefore, here, the FRU is the EIF mentioned above
- Determined by referring to the FRU correspondence table, these FRU
The information specified by RU'i is displayed on the console 7 or printed out on the printer 8.

Next, in Figure 4 G, the processing contrast is EIF in the following order.
and if all EIF checks are finished,
If not (Figure 4), this updated E-I
F becomes a new object of processing, and the above-described processing is repeated from FIG.

By the above processing, the original EIF can be created without any human intervention.
is automatically determined, and each FR corresponding to this primitive EIF
The information pointing to U is from console 7 or printer 8.
The diagnostic process is quickly and automatically performed.

The above shows one embodiment of the present invention, and the present invention is not limited to the above embodiment.

and <VC Figure 3 (A), @, <avc shows 1EIF
The set/reset table, the error propagation information table, and the EIF-FRU correspondence table are merely examples.

Similarly, the flowchart shown in FIG. 4 also shows one example, and it is clear that the present invention is not limited to the above embodiment.

〔Effect of the invention〕

As described above, according to the present invention, by preparing information regarding error propagation VC in advance and referring to this information,
A source EIF set directly by an actual hardware failure is set by a propagated EIF due to error propagation.
It is possible to provide a diagnostic processing method that analyzes the error log information hidden by F, determines the original EIFt, and automatically outputs the corresponding maintenance/replacement unit (FRU) information without human intervention. .

This streamlines the diagnostic process and significantly reduces the effort and time required.

[Brief explanation of the drawing]

FIG. 1 is a block diagram for explaining one embodiment of the present invention, FIG. 2(A) is a diagram showing an example of the arrangement of an error detection circuit and an EIF in each device, and FIG. The transmission % for the original EIF and its n in the example shown in Figure 2
The EIF table is shown in Figure 2.
Figure 3 (A shows an example of an EIF set/reset table, Figure 3 1 (Bl shows an example of an error propagation information table), Figure 31Bl (C) shows an example of a 1-IEIF-IU correspondence table. Figure 4 is a diagram showing a 70-chart showing the operation of this implementation.In the figure, 1... Arithmetic processing unit (EPU),
2...System control unit (8CU), 3...
... Input/output control unit (IOP), 4... Main memory unit (MMU), 5... Service processor (SVP), 6... Disk, 7...・・・・・・
Console, 8...Printer, D1-D6...
...Error detection times JEIFI~E I F4...
...Error indicator flip-flop (ELF)
), Pi~P6...Process $3

Claims (1)

[Claims] A plurality of hardware error detection circuits, an error indicator flip-flop (hereinafter abbreviated as EIF) corresponding to the error detection circuit that is set in response to error detection by each of the error detection circuits, and an error detection circuit. an error log collecting means for collecting an error log including information indicating the set/reset state of each EIF when an error occurs, and analyzing information indicating the set/reset state of each EIF in the collected error log. In a diagnostic processing method that points out a failure location, the EIF that is initially set when an error occurs is defined as the primitive EIF, and the EIF that is set upon detection of an error that propagated this error is defined as the primitive EL.
When defined as a propagated EIF for F, for each EIF, prepare error propagation information that makes all the EIFs that can become the original EIF such that the EIF can be the propagated EIF as an upper EIF for the EIF, Each of the EIFs set in the information indicating the set/reset status of each EIF in the collected error log
For each IF, if the upper EIFs for the EIF in the error propagation information have all been reset in the error log, the EIF is determined to be the original EIF and the failure location corresponding to the EIF is pointed out. A diagnostic processing method characterized by: