JPS60171544A - Self-diagnosis device for abnormality of computer system - Google Patents

Abstract

PURPOSE:To suppress a computer system to minimize the discontinuation of the function by detecting a fault factor of the software by the computer system itself and removing said fault factor. CONSTITUTION:A level action diagnosis module 2 and an action monitor timer 3 are provided for each working level of all tasks for an operating system OS1. The module 2 is started periodically by the OS1 and sets the fixed time value to a timer 3. The timer 3 sets 1 to a level action monitor register 4 after the count-up of time to store an abnormality. The OS1 monitors the working time of each task and sets this time value to a tast working time memory 5. A system diagnosis device 6 works periodically to analyze the contents of a fault and detects the direct factor of the abnormality. Then the device 6 delivers a stop request of the corresponding tast to the OS1 and also displays the abnormality contents to an external abnormality display device 7.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to an abnormality diagnosis device for a computer system, and in particular,
The present invention relates to a self-diagnosis device that suppresses a total system outage caused by a software defect to a minimum local outage.

[Background of the invention]

An operating system (hereinafter referred to as 0
S) In a computer system with t-, the most important point is to always execute all processing functions without delay.

Conventionally, the O8 has determined most of the program defects by itself, and has been successful to some extent in minimizing the spread of the defects to the entire system.

For example, when a defect is discovered during a data write area protection check or invalid instruction execution check, this function immediately disables the corresponding task (hereinafter the smallest unit of an independently operating program is referred to as a task). Applies to.

However, defects that have been difficult to suppress until now,
There was an endless loop of the program and a tend lock between tasks due to commands related to possession and release to prevent shared resources (data files) from being destroyed. In computer systems, tasks are given several action levels, and the importance and importance of processing are determined.
Priority is given according to the level of urgency, but if a task at a certain level falls into an eternal loop, all tasks at a lower level will be placed in a waiting state and will essentially stop functioning. Furthermore, if an occupancy/WC removal command is issued incorrectly, all functions for inputting and outputting to a certain resource are stopped. Thus, these defects are likely to cause the entire system to stop functioning.

To deal with this, various methods have been considered and have not been put into practical use.
I've been doing it. For example, some timers may be used to detect endless loops. This is a method in which the maximum processing time is calculated and set in advance for each task, and when this time is exceeded, the execution of all tasks is prohibited. This method provides a perfect countermeasure for the task in question, but if the task is not the direct cause, then the wrong action has been taken and the situation will worsen. Furthermore, since the setting of the maximum processing time is left to the designer of each task, in many cases it has not been sufficiently considered or set, making it unreliable. In addition, as a method for the entire system, a task is set at the highest level. There is a method of turning the system on, but this method only detects an abnormality and is not effective in preventing system stoppage.

As described above, it has been impossible to achieve sufficient effects with conventional methods.

[Purpose of the invention]

An object of the present invention is to provide a device that minimizes system outages by systematically identifying the causes of software failures in a short time and accurately, and directly removing the causes by the system itself. .

[Summary of the invention]

The present invention is characterized in that it does not depend on the designer of each program to set the abnormality detection function, and the computer system itself performs all the detection and removal of the cause. That is, O8
periodically diagnoses whether the operation is possible or not for each level,
Moreover, the purpose is to constantly measure the processing time of all tasks, determine the cause of the abnormality from these two pieces of information, and remove the relevant cause.

[Embodiments of the invention]

An embodiment of the present invention will be described below with reference to FIGS. 1 to 4.

Figure 1 shows the effects on the system when one task falls into an endless loop.

The vertical axis shows the task operation level (set to level 7 in this example), and the horizontal axis shows the elapsed time. In the figure, time t. ~1
o, the system is operating normally, and the higher levels (levels with lower numbers are referred to as higher levels, 1., L2..
...Task L7) is operating with priority. t8~t
9, task ■ occupies a certain resource (referred to as reserve [F]). Task (2) starts operating from to, but due to some reason, it becomes in an endless loop state. After that, the high-level task (■) operates normally, but the low-level tasks (2) to (2) enter a waiting state and stop functioning. Also, if a high-level task also tries to occupy that resource,
It will be waiting for release and will stop functioning. '2
The entire system will stop functioning.

FIG. 2 shows a configuration example of a computer system abnormality self-diagnosis device.

Under the main cover 081, a level operation diagnosis module 2 and an operation monitoring timer 3 are provided for each task operation level,
2 is periodically activated by 1 and sets all values to 3 for a certain period of time. When 3 is a time amplifier, 1 is sent to the level movement 11 monitoring V register 4, and an abnormality is memorized. O8 monitors the operation time of each task and sets this value in the task operation time memory 5. The system diagnostic device 6 operates periodically, analyzes the contents of the abnormality, searches for the direct cause, issues a request to the IVc to stop the corresponding task, and displays the contents of the abnormality on the external abnormality display device 7. do.

Here, 2 is configured to perform an operation equivalent to a task and is connected to an operation queue. Therefore, each level of operation is detected using the three metals.

FIG. 3 shows the detailed configuration of 6. 6 is the 4.5 meeting input. First, it is determined whether 4 is all zero or not, that is, whether there is an abnormality in the system or not. Next, the flag O of 4 is determined.
After all abnormal levels are detected by the leftmost N number determining device 68, the direct cause is analyzed by the abnormal cause task determining device 69vc.

69 extracts the operating times of the tasks included in the abnormality level determined in 68 from 5, and determines the maximum among them as the direct cause. After that, task operation stop request device @:610
The abnormality display device 612 displays and outputs a stop request 611 to the outside.

FIG. 4 shows an example in which all of the conditions shown in FIG. 1 are suppressed with minimal influence by this device.

'! At 1, the entire system is temporarily stopped, but when C1 to C7 operate, the c1 to Cuff level is high, so they operate, and at 3, the timer hall all cents, but C4 to C
7 cannot set the value to 3 due to the eternal loop of task ■. 3 is t, time amplification is performed at 2, 4 is turned on, 6 detects task 2, and 1 disables task 2. As a result, the resource-occupying task ■ operates and is released (q
c) Do.

As a result, the tasks in the wait state are released from the wait state one after another, and at t24, all functions other than ■ in the system become normal, and the spread of the abnormality can be minimized.

〔Effect of the invention〕

According to the present invention, since the computer itself can detect software abnormalities and eliminate the cause thereof, the reliability and operation path of the computer system can be improved.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram of system l and edge spread due to the eternal task loop of the present invention, FIG. 2 is a block diagram of the computer system abnormality self-diagnosis device of the present invention, and FIG. 3 is a detailed 1 configuration diagram of the present invention. FIG. 4 is a side view of the abnormality suppression operation of the present invention.

Claims (1)

[Claims] 1. In a computer system having an operating system in which a plurality of independent programs operate simultaneously and control them according to a fixed procedure, a level operation diagnostic device is provided for each operation level of the program and monitors the operation, and an operation monitor. A timer, a level operation monitoring register to which data is set when the operation monitoring timer times up, this level operation monitoring register and program operation time memory are input, analyze the cause of system abnormality, remove all applicable causes, A computer system abnormality self-diagnosis device comprising a system diagnosis device that displays an external display.