JPS598064A - Fault diagnosing system for multiplex computer system - Google Patents

Abstract

PURPOSE:To prevent the loss of an important key for fault diagnosis of a multiplex computer, by collecting the information existing on a main memory of the series having a fault by means of remaining nondefective series. CONSTITUTION:If a fault arises to a computer of a series (a), the generation of the fault is informed to a CPU1b of a series (b). Receiving the information on the fault discontinuation of the series (a), a main memory information shunting program 2b-2 of the series (b) keeps the information existing on a main memory 2a of the series (a) in a main memory information preserving region 3b-2 of an external storage device (bulk memory) 3b of the series (b) via a signal system C and devices 5a and 5b. A CPU1a can give an access a main memory 2b of the remote system via the devices 5a and 5b. While the CPU1b can give an access to the memory 2a of the remote system via the devices 5b and 5a respectively.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a fault diagnosis method for a multi-system computer system, and particularly to a fault diagnosis method for a multi-system computer system that can save information up to just before the fault occurs in the main memory of a computer in which a fault has occurred. This paper relates to a method for diagnosing problems in computer systems.

[Technical evidence of the invention]

In general, immediately before the occurrence of a serious failure that causes a computer system to stop, failure of an important part of the hardware constituting the computer system or M-running of a program due to a bug can be considered.

The most effective clue information for diagnosing these failures and determining the cause of the failure exists in the main memory of the subtraction machine system at the time of the failure. The first reason is that the information stored in the main memory when the program stops includes information such as how the program was running until the program stopped and the input/output status with external storage devices and peripheral devices. For this reason,
Conventionally, when a machine comes to a stop, the information in the main memory is temporarily evacuated to an external storage device, and after the training machine system is restarted, that information is output to a line printer, etc., and the fault is diagnosed. There is.

A conventional fault diagnosis method will be explained with reference to FIG. 1st
The computer system shown in the figure is a central processing unit (hereinafter referred to as CP).
(hereinafter referred to as U) 1, main memory 2, external storage device (hereinafter referred to as bulk memory) 3, line printer (hereinafter referred to as LP)
There are 4. (with 6 passes).

Now, when a failure occurs in this machine system due to the above-mentioned hardware or software, it is sent to the CPU in the form of a normal interrupt (hereinafter referred to as failure interrupt).
1 will be notified. C that received the failure interrupt here
PU 1 interrupts the program that was running until then, and
, immediately move the side leg to the main memory information saving program 2-1. The main memory information saving program 2-1 is a program that needs to run under these circumstances, that is, just before the subtraction machine system is about to stop, so it is always saved in bulk memory like a normal program. 3. Reeds on top,
It is not possible to load it into main memory 2 only during execution,
This is a program that resides in main memory (referred to as a main resident program).

Since the operation of the main memory information saving program 2-1 is well known, a detailed explanation will be omitted, but it has the following functions.

That is, the information stored in the main memory 2 is transferred either completely or selectively to the main memory information storage area 3 of the bulk memory 3 through the signal system A.
-1 and then stop the 61 computer system. At 7, main memory information storage area 3- of bulk memory 3
Saved in 1. After restarting the subtractor system, the information shown in
It is output to the LP4 hostel for troubleshooting.

[Problems with countermeasure technology]

The above is a typical fault diagnosis method in a computer system [F
However, it has the following drawbacks7. In other words, if the cause of the failure is in the hardware and it becomes impossible to generate the failure interrupt, or if the signal system AVrC becomes impossible to transfer to the bulk memory 3, this method will no longer function at all. That's what I'm saying.

Furthermore, the same is true when the fault lies in software and the main memory information saving program 2-1 is destroyed due to program failure.

A multisystem system also obtains information for fault diagnosis using the same method as above.

[Purpose of the invention]

The present invention has been made for the purpose of solving the above-mentioned drawbacks, and provides a fault diagnosis for a multi-redundant machine system that can prevent the loss of important clues for fault diagnosis even when a fault occurs in either hardware or software. The purpose is to provide a method.

[Summary of the invention]

In the present invention, when a failure occurs in any of the computer systems that make up a multi-system, information stored in the main memory of the computer system in which the failure has occurred is collected from the remaining normally operating systems. This is intended to prevent the loss of important clues for diagnosis.

Embodiments Hereinafter, embodiments will be described with reference to the drawings. FIG. 2 is a block diagram of one implementation of the fault diagnosis method for a multi-system computer system according to the present invention.

FIG. 2 shows a dual computer system, each of which has CPUs 1a and 1b and a main memory 2a.

2b, bulk memory 3a+3bx LP4a, 4b are provided as in the 11th ¥1. Please note that the suffix a is 1. We will call the machines that have been reduced to 1st series, and the 2nd series that has b attached to them.

5a and 5b are FIs for allowing each other to access the main memory of the other system, that is, the CPU 1a i7tgW
5 Main memo IJ in the other party's system via B+5b 2
The CPU 1b can access the device 5b.

The main memo 'J2a in the other system can be accessed via the computer system 5a (also σ), which will be referred to as a computer system linkage device (hereinafter referred to as C8L).

Next, the operation of the example of the rudder shown in FIG. 2 will be explained using the flow goo yard vC of FIG.

Now, to explain the case where a failure occurs in the old computer of the first series, due to the OR condition of steps A and B, step fCV?
The CPU 2 b of the second series indicated by X is notified of the occurrence of the failure. That is, the stop detection device W1 of the first series (not shown)
．． Either by software means such as inputting the output to the interrupt handling device of the second series (step A), or by software means such as detection by the other system status monitoring program in the second series (step B). Depending on the
The normal second system learns of the occurrence of a failure in the first system, 2- (Step C).

The main memory information saving program 2b-2 of the second series, which has been notified of the failure stop of the first series, saves the information in the main memory 21 of the first series that has stopped due to the failure via the signal system C to C8L.
5a and 5b, and is stored in the main memory information storage area 3b-2 of the second series bulk memo I73b (step D).

Note that the above-mentioned operation of the main memory information saving program 2b-2 in the second line of management system can be performed in parallel with the execution of other business programs.

Note that the computers constituting the multi-system computer system do not operate independently, but the η-resistant machines operate in an organically coupled manner. Therefore, the main memory save program 2b-2 of the second series explained in the above embodiment collects the information stored in the main memory of the first series, and
If the information on the main memory 2b of the series itself is also stored in the main memory information storage areas 3b-2 in the bulk memory 3bvC, it is possible to provide information for a wider range of fault diagnosis. .

〔Effect of the invention〕

As described above, according to the present invention, multiple series words 1
In a computer system, when a failure occurs in any of the computer systems that make up the 11 systems, information on the main 2 memory of the system in which the failure occurred is collected from the remaining normally operating systems. , if necessary, it is also possible to collect +9 information at the same time related to the state of the faulty series in the main memory of the normal system 1 (σ)
Therefore, it is possible to provide a fault diagnosis method for a multi-system system that is more accurate and does not cause loss of information for extensive fault diagnosis.

[Brief explanation of the drawing]

FIG. 1 is a configuration diagram for explaining a conventional fault diagnosis method.
FIG. 2d is a block diagram for explaining a fault diagnosis method for a multi-system computer system according to the present invention, and FIG. 3 is a flowchart for explaining the operation. ■...Central processing unit M2...Main memory 2-1
. . . Main memory information saving program 3 . . . External storage device 3-1 . . . Main memory information storage area 4 . . . Line printers 5a, 5b . Person Tokyo Shibaura Electric Co., Ltd.

Claims (1)

[Claims] In a fault diagnosis method for a multi-computer system that can save information in the main memory of the faulty computer without losing it when a fault occurs in a multi-computer system consisting of multiple computers, a main system that operates when a fault occurs is used. A fault diagnosis method for a multi-system subtractor system characterized by collecting information in the main memory of a faulty computer into the main memory information storage area of a normally operating computer using a memory saving program.