JPH05257675A - Program parts duplex system - Google Patents

Abstract

PURPOSE:To improve reliability of the system by constituting the system so that the operation can be continued by switching other program parts having the same function, in the case where the on-line system becomes abnormal being caused by a software. CONSTITUTION:In a fault recovery processing of an on-line system, the system is provided with a means 14 for managing a program parts group in a storage area, a means 12 for collecting fault statistical information for every program parts in which a fault is occurred, a means 13 for selecting a program whose fault occurrence frequency is less from the program parts group having the same function as the program parts which cause a fault and switching to the active program, and a means 16 for scheduling the active program parts and shifting the control, and by switching the program parts concerned at the time of fault of the software, the operation of the system is continued.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to improving software reliability in an online system, and more particularly to an online program parts duplication system.

[0002]

2. Description of the Related Art Conventionally, in an online system, when a failure occurs during service, the system is down if it is impossible to continue operation. In this case, even if the cause of the failure is software, the software was restarted, the same software was stored in the memory, and the operation was restarted.

[0003]

When the above-described conventional online system is restarted from system down, the same software is stored in the memory and the operation is restarted. However, since the same software operates, there is a drawback that if the software has a cause of failure, the system goes down again when data that causes the failure is input.

[0004]

The program component duplication system of the present invention provides means for managing a group of program components in a storage area and fault statistical information for each faulted program component in a failure recovery process of an online system. A means for collecting, a means for selecting a program with a low failure frequency from the program parts group having the same function as the failed program part and switching to the active program, and scheduling the active program part to transfer control. And a means for doing so to switch the corresponding program component at the time of software failure and continue the operation of the system.

[0005]

The present invention will be described below with reference to the drawings.

FIG. 1 is an explanatory view showing an embodiment of the present invention. In the same figure, the program component duplication system according to the present invention uses an OS (operating system) fault processing unit 11, fault statistical information collecting unit 12, program component switching control unit 13, program component information management table unit 14, various program groups 15, Program parts schedule section 1
Including 6. It is assumed that the program group 15 and the program component information management table 14 are stored in the memory by the OS when the system is started up.

[0007] While a program is running, an exception interrupt for writing data to an area to which real memory is not allocated occurs. When a failure due to such software occurs, control is transferred to the failure processing unit 11 of the OS.

In response to this, the OS failure processing unit 11 performs failure analysis and recovery processing. At this time, if the fault content is a fault caused by software, the fault statistical information collecting unit 12
Notify the occurrence of a failure to.

The fault information collection unit 12 judges whether a fault has occurred in the duplicated program component prior to the collection of the fault statistical information. When running in a program part, the corresponding part group I at the start of the program part
D and the part ID are set in the running part identification ID save area on the memory, and the running part identification ID save area is cleared at the end of running. Therefore, if the running part identification ID save area is not cleared, it is determined that a failure has occurred in the program part.

If the running part identification ID save area is cleared, it is not a duplicated program part, so that switching of the program part is not necessary and control is returned to the failure processing section 11 of the OS.

When it is determined that a failure has occurred in the duplicated program component, the failure statistical information collecting unit 12
Parts ID set in the running parts identification ID save area
And the component ID, the program component information management table of the program component information management table unit 14 is referred to, and a table entry corresponding to the component group ID and the component ID is obtained.

FIG. 2 is an explanatory diagram showing an example of the program component information management table. In the figure, the shaded portion represents one entry in the table.

Next, the statistical information is collected in the corresponding entry of the program component information management table. In this embodiment, the number of failure occurrences and the addition of the continuous failure occurrence counter are added. The continuous failure means an event in which the same part is selected twice in a row and becomes a failure twice in a row. When the collection of the failure statistical information is completed, the control is transferred to the program part switching control unit 13 to switch the program parts.

FIG. 3 is a flow chart showing the operation of the program switching control unit 13. The input information of the program component switching control unit 13 is the component group ID and the component ID.

(Step 21): Obtain all entries corresponding to the component group ID from the program component information management table 14 on the memory. In this embodiment, one component group ID has two components, so two entries are obtained.

(Step 22): Next, the number of fault occurrences in the fault statistical information area is compared for all entries corresponding to parts, and the entry corresponding to the part having the smallest value is selected. If there are the same number of failures, the entry corresponding to the first detected component is preferentially selected.

(Step 23): The continuous failure occurrence counter of the selected component-corresponding entry is referred to. If it is smaller than 2, the process proceeds to step 26.

(Step 24): The continuous failure occurrence counter is cleared.

(Step 25): Obtain an entry corresponding to the other duplicated part.

(Step 26): In the program parts information management table 14, the working / spare display of the entry corresponding to the faulty part is set as the spare, and the working / spare display of the entry corresponding to the newly selected part is set as the working.

The switching of the program component in which the fault has occurred is completed, and the process returns to the fault processing unit 11 of the OS.

Next, how the actually changed program parts run will be described. When it is necessary to use a function that has been made into a program component while a program in the various program groups 15 is running,
The component group ID of the target program is specified, and the program component scheduling unit 16 is requested to start the program component.

In response to this, the program parts schedule section 16 refers to the program parts information management table held by the program parts information management table section 14, and selects the active / active entry from all the parts correspondence entries corresponding to the specified parts group ID. The part corresponding entry whose preliminary display is currently used is selected, the program entry address in the corresponding entry is obtained, and the control is transferred. In this way, the program component can be operated.

The present invention is not limited to the case where the program parts having the same function are duplicated, and it is possible to have three or more program parts.

[0025]

As described above, according to the present invention, when the online system becomes abnormal due to software,
Since it is possible to continue operation by switching to another program part having the same function, there is an effect that the reliability of the system is greatly improved.

[Brief description of drawings]

FIG. 1 is an explanatory view showing an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing an example of a program component information management table.

FIG. 3 is a flowchart showing an operation of a program parts switching control unit.

[Explanation of symbols]

 11 OS Failure Processing Section 12 Failure Statistics Information Collection Section 13 Program Parts Switching Control Section 14 Program Parts Information Management Table Section 15 Various Programs 16 Program Parts Scheduling Section

Claims (1)

[Claims] 1. In a failure recovery process of an online system, a means for managing a program part group in a storage area, a means for collecting failure statistical information for each failed program part, and the same program part as the failure A program part corresponding to a software failure is provided with means for selecting a program having a less frequent occurrence of a failure from the program parts group having functions and switching it to the active part, and means for scheduling the active program part and transferring control. The program component duplication method characterized by continuing to operate the system by switching between the two.