JPH04148341A - Task recovery system - Google Patents

Abstract

PURPOSE:To localize the faulty wave and range of an online system by operating the garbage collection of a task at which a fault occurs. CONSTITUTION:This system is equipped with a task initialization setting part 13 which operates an interruption register for the transition of a control when the fault occurs at the task in a center, and a task recovery part 14 which operates the garbage collection of the task in which the fault occurs. Also an abort library part 15 which calls the task recovery part 14 when the fault occurs at the task, and a representative task part 16 which prepares the task again are provided. The task recovery part 14 operates the garbage collection as the processings such as the return of a memory kept by the task at which the fault occurs, the close of a file, and a lock release, and then the representative task part 16 prepares the task again. Thus, only the pertinent task can be saved without interrupting the online system, and the faulty wave and range of the online system can be localized.

Description

[Detailed Description of the Invention] [Summary] Regarding a task recovery method in a personal computer communication system, an object of the present invention is to provide a task recovery method that can localize the scope of failure in an online system by performing garbage collection. In a communication system that performs communication between a center and a terminal device, the center includes a task initial setting unit that registers an interrupt for transferring control when a failure occurs in a task, and a task initialization unit that registers an interrupt for transferring control when a failure occurs in a task. A task recovery section that performs garbage collection, an aport library section that calls the task recovery section when a failure occurs in a task,
It is equipped with a representative task section that regenerates tasks, and is configured to rescue only the relevant tasks without interrupting the online system.

[Industrial Application Field] The present invention relates to a task recovery method in a personal computer communication system.

Currently, it is said that there are approximately several hundred thousand PC communication users in Japan, and this number is expected to increase in the future. Therefore, if the online system is disrupted and the service is interrupted, a large number of users will be affected. Therefore, it is necessary to devise ways to prevent a partial task failure from affecting the entire online system.

[Prior Art] As a conventional personal computer communication system, there is one shown in FIG. 5, for example.

In FIG. 5, 1.2 is a terminal device, 3 is a center, and the terminal devices 1 and 2 are connected to the center 3 via a line network 4. In FIG. The line network 4 has a large number of access points 5, and a large number of terminal devices 1-2 perform personal computer communication with tasks (application tasks) 7 within the center 3 via communication paths 6 from the access points 5. Receive various services.

In such conventional PC communication systems, if a failure occurs for some reason in task 7 that controls terminal device 1.2, forced logout processing (mainly disconnection of communication path 6) is performed. Ta.

However, the various resources used by Task 7 (
(files, memory, etc.) (processing such as returning memory, closing files, releasing locks, etc., hereinafter referred to as garbage collection) was not performed.

[Problem to be solved by the invention] In such a conventional PC communication system, when a failure occurs in a task, logout processing is performed, but garbage collection is not performed, so various resources remain. This causes inconsistency in subsequent system operations, and as a result, there is a problem in that the entire online system may stop.

The present invention has been made in view of such conventional problems, and an object of the present invention is to provide a task recovery method that can localize the scope of failure in an online system by performing garbage collection. It is said that

[Means to solve the problem] FIG. 1 is a diagram explaining the principle of the present invention.

In the 1111th task, 13 is a task initial setting unit that performs interrupt registration for transferring control when a failure occurs in task 11, 14 is a task recovery unit that performs garbage collection for task 11 in which a failure has occurred, and 15 is a task An abort library section 16 calls the task recovery section 14 when a failure occurs in the task 11, and a representative task section 16 generates the task 11 again.

[Operations] In the present invention, first, when a failure occurs in a task, the task initialization unit registers an interrupt to transfer control to the task recovery unit, and when a failure occurs in a task, the abort library unit registers an interrupt. Transfers control to and calls the task recovery section.

The task recovery section performs garbage collection to return memory secured by the task, close files, release locks, etc., and then the representative task section generates a task again.

In this way, if a failure occurs in a task and it becomes impossible to continue processing, only the relevant task can be rescued without interrupting the online system, thereby localizing the scope of the failure in the online system. be able to.

[Example code] Embodiments of the present invention will be described below based on the drawings.

FIGS. 2 to 4 are diagrams showing one embodiment of the present invention.

In FIG. 2, reference numeral 11 indicates a center-side task (application task) for a user-side terminal device to communicate with a personal computer and obtain a predetermined service.

Task 11 detects a failure due to data inconsistency,
There are two cases: a case in which the task 11 itself detects a fault due to a program bug, and a case in which an interrupt occurs due to execution of an undefined instruction, memory protection, etc. (hereinafter referred to as a signal). In task recovery, it is necessary to somehow recognize these two cases and seize control. Therefore, a method for recognizing the former case is to provide a dedicated library for calling task recovery, and a method for recognizing the latter case is to intercept signals.

Reference numeral 13 denotes a task initial setting unit, and in response to a task initial setting call from the task 11, the task initial setting unit 13 registers a signal for transferring control to the task recovery unit 14 when a failure occurs in the task 11. . Signal management information for signal registration is stored in advance in a shared space in the memory 12 that can be referenced by each task. As shown in Figure 3, the signal management information consists of a signal number, signal setting value, task relief display, and backup.
It is set for the number of tasks. The signal numbers are 1: access exception, 2: illegal instruction, 3 nihard failure, etc.
Signal setting values are: n: No signal registration, d
i: Disable the signal, a near-bort, t: Call task recovery, etc., and the task relief display is y: rescue, n: do not rescue.

Reference numeral 15 denotes an abort library section. When a failure occurs in the task 11, control is transferred to the abort library section 15, from which the task recovery section 14 is called.

The task recovery unit 14 checks whether the task 11 in which a failure has occurred is a target for relief. The tasks 11 to be rescued are limited to those that are unlikely to cause trouble to the entire system from now on. - For boat fishing, there is a task other than the representative task unit 16 that controls the system, and it can be determined arbitrarily (see signal management information). Task recovery part 1
4 performs garbage collection such as returning various memories, closing various files, and releasing locks.

16 is a representative task unit, and when the representative task unit 16 receives a failure notification of task 11 from the task recovery unit 14,
The program is reloaded from the program file 17 and the task 18 is generated again.

Next, the operation will be explained.

FIG. 4 is a flowchart illustrating the present invention in detail.

In FIG. 4, first, in step S1, in order to set the intercept destination at the time of signal generation to the task recovery unit 14,
The task initial setting unit 13 performs signal registration (see FIG. 3).

Next, if a failure occurs in the task 11 in step S2 (the two cases described above), control is passed to the abort library unit 15 in step S3, and the task recovery unit 14 is called.

The task recovery unit 14 locks the routine in step S4, sets the signal again in step S5, and determines whether the task recovery is running in step S6.

If the vehicle is not running, a message indicating that a task failure has occurred is output in step S7, and if the vehicle is running, the process advances to step S9.

Next, in step S8, it is determined whether or not the task 11 is to be rescued, and if it is the task 11 to be rescued, a failure notification is sent to the representative task unit IG in step SIO.
If the task 11 is not the relief target, the process advances to step S9.

Next, in step S11, various memories secured by the task 11 are returned, and in step 812, various files secured by the task 11 are closed. Then, in step S13, the locked resources are forcibly released, and in step S14, memory information necessary for easily troubleshooting is acquired.

Next, in step 815, the routine lock is released,
In step 316, task 11 is deleted.

Thereafter, in step 817, the representative task unit 16 reloads the program and regenerates the task 18, taking into account the damage to the internal data section and program text section that the failed task 11 had.

In this way, without interrupting your online system,
Only the relevant task 11 can be rescued, and the extent of the failure in the online system can be localized.

[Effects of the Invention] As explained above, according to the present invention, when a failure occurs in a task and processing cannot be continued, only the relevant task can be rescued without interrupting the online system. This makes it possible to localize the range of failures in online systems.

[Brief explanation of the drawing]

Fig. 1 is a diagram explaining the principle of the present invention, Fig. 2 is a diagram showing an embodiment of the invention, Fig. 3 is a diagram showing signal management information, Fig. 4 is a flowchart explaining the operation, and Fig. 5 is a diagram showing an embodiment of the invention. It is a figure showing a conventional example. In the figure, 11.18...Task, 12...Memory, 13...Task initial setting section, 14...Task recovery section, 15...Abort library section, 16...Representative task section, 17...Program file.

Claims (1)

[Claims] In a communication system that performs communication between a center and a terminal device, there is provided a task initial setting for registering an interrupt in the center to transfer control when a failure occurs in a task (11). (13) and a task recovery unit (14) that performs garbage collection of the task (11) in which a failure has occurred.
and an abort library unit (15) that calls the task recovery unit (14) when a failure occurs in the task (11).
) and the representative task part (1) that regenerates task (11).
6), the corresponding task (
11) A task recovery method characterized by relieving only.