JPH08249201A - Data processor - Google Patents

Abstract

PURPOSE: To provide a data processor capable of reducing the definition amount of job control and easily adding and eliminating jobs by using plural job control tables for defining the control information of the jobs for every job and retrieving the control information corresponding to an event. CONSTITUTION: For state transition charts 121 and 122. the state of the job at present is taken in a row and the event is taken in a column and processings and new job states under respective conditions are defined. For priority charts 111 and 112, computers are arranged in the descending order of priority. In a normal state, the computer A is an operating system and executes the job 1 and the job 2. When the computer A is stopped, the computer B takes over the jobs 1 and 2. Also, an operating state table 101 is provided so as to manage the jobs executed by the respective computers within a system. Entries for the number of the computers are provided in the table 101 and whether the computer A or the computer B is operated and which job is executed are managed respectively.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is defined in advance in a data processing device for simultaneously executing a plurality of jobs when the execution of a job is stopped due to some cause (whether hardware trouble or software trouble). The present invention relates to a data processing device that recovers based on the control information and resumes the execution of work.

[0002]

2. Description of the Related Art In a dual system, when one computer stops, the other one takes over the work,
Configure a non-stop computer system. Since the work being executed on each computer changes due to the inheritance of this work, it is not possible to specify which computer is executing which work at a certain time. The current state depends on the order in which events such as computer shutdown occurred in the past. In addition, in the business takeover process, the business that was being executed on the stopped computer was interrupted during execution, and data integrity may have been lost. Have to start again with. At this time, the recovery process to be executed differs depending on which computer has stopped or which job has stopped.

Therefore, in what case the current state (state of each computer and execution state of each job), what kind of event (stop of one computer, etc.), what kind of processing, what It is necessary to have a definition of whether to move to such a state. A state transition diagram as shown in FIG. 16 is generally used as a means for expressing this so that it is easy to understand. (However, in FIG. 16, the processing when an event occurs is omitted.)

FIG. 16 is a state transition diagram showing a dual hot standby system for two jobs. In this figure, each ellipse (for example, 200) shows the state of the system, the upper part of the ellipse shows the current state of the computer A, and the lower part of the ellipse shows the current state of the computer B. This means that when the event attached to each arrow (for example, 202) occurs, the state transits to the state indicated by the arrow. For example, when the computer A is stopped in the state 200 in which the tasks 1 and 2 are executed on the computer A, the system state transits to 201 in which the tasks 1 and 2 are restored on the computer B. (Note: It is described here that the stopped computer will not be recovered before the business recovery processing is completed. This is because even in an actual system, the cause of the computer stop must be removed. The state transition diagram (table) becomes more complicated if such a case is taken into consideration.) However, since the information in the diagram is difficult to handle with a computer, this Is converted into an equivalent table (= state transition table: FIG. 17), and the system definition is given to the computer as a table. Making the state transition diagram an equivalent table is also effective in preventing system state omissions (definition omissions). FIG. 17 shows a current system state in a row and an event in a column. When a certain system state occurs, which event occurs, what kind of processing is performed, and which system state is transitioned to. (However, similar to FIG. 16, the processing at the time of event occurrence is omitted in FIG. 17 as well.)

In FIG. 17, the upper part of the system state shown in the column indicated by 202 shows the state of the computer A, and the lower part shows the computer B.
Indicates the state of. The number 203 assigned to each system state is used when designating a new system state. Each element of the table of FIG. 17 has an event (for example, 204) of the column in the system state of the row.
Defines a new system state (the system state with the number shown after the arrow in the figure) in the case of occurrence, and the hatched line means that the combination of the system state and the event does not occur. For example, in the state where tasks 1 and 2 are being executed on computer A (system state number 4),
When the computer A is stopped, the system state changes to the system state number 11 for recovering the tasks 1 and 2 on the computer B.

Even in a simple system as shown in FIGS. 16 and 17, the number of system states and the number of events are relatively large.
Therefore, the state transition table also becomes large. However, in reality, there are many combinations of states and events that do not occur, which wastes memory space. Further, the number of system states increases as the number of tasks increases, which increases the state transition table and the memory space therefor.

[0007] In such a hot standby system, one is clearly divided into an operating system and the other is a standby system (backup machine), so the number of states is still small, but the loads that take over the work of each other. In the share system, each work must be considered independently, and the number of states and the number of events increase even if the number of works is the same.

FIG. 18 shows a state transition diagram of the dual load sharing system for two jobs, and FIG. 19 shows a state transition table equivalent to FIG. In the load sharing system, there are possible states 205/206 (same as hot standby) that are concentrated on one computer, and states 207/208 that work is being executed on each computer. 2 more than hot standby. Further, instead of switching the operating system and standby system like hot standby, it is possible to change the computer that executes each task, so the number of events increases accordingly. As a result, in the dual load sharing system, the number of system states and the number of events roughly follow the formulas below. In the following equation, N represents the number of jobs and nCi represents the number of combinations from n to i.

[Equation 1]

The number of system states is three times the total number of combinations because both the combination of each business and the business suspension / recovery processing due to the suspension of each computer in each combination occur, and both are suspended. It becomes +1 for the state. Regarding the number of events, for each movement of work,
It is +5 for the start and stop of each computer and the initial start of the system. As described above, the number of states and the number of events respectively depend on the number of jobs. Therefore, the state transition table defined by the combination of the states and the events has a rate of more than the square of the number of jobs as the number of jobs increases. To increase.

For example, FIG. 20 shows the state of the system when there is one computer and the number of tasks is three. As shown in FIG. 20, the number of system states at this time is eight. On the other hand, FIG. 21 shows the state of the system when there are two computers and the number of jobs is three. As shown in FIG. 21, the number of states of the system at this time increases to 27. Also, since the information that must be defined as the management information 16 is the number of {system state} × {event} number, the state transition table of the system in the case shown in FIG. 21 becomes very complicated, Since there are many numbers, they are not shown here.

As described above, in the conventional method, since the state transition table, which is the definition information of the system, greatly increases with the increase in the number of jobs, it is impossible to increase the jobs that can be controlled independently. This means that a plurality of originally independent jobs must be handled collectively in the system definition.

In a multi-system, an increase in the number of computers causes an increase in the state transition table, making definition difficult and consuming memory space more serious. In the multiple load sharing system, the number of system states and the number of events roughly follow the formulas below. This equation is a generalization of the equation in the above-mentioned dual system to a multiple system. In the following formula, N represents the number of jobs, M represents the number of computers, and nCi represents the number of combinations from n to i.

[Equation 2] The number of system states is a complicated formula as described above, because there is a combination of jobs in each computer, and there is a job stop / recovery process by stopping each computer. The number of events will be + ({number of computers} × 2 + 1) for the start and stop of each computer and the initial startup of the system in addition to the movement of each business.

[0013]

In the conventional data processing apparatus as described above, it is necessary to redefine (replace) the state transition table when adding a new job or a new computer to the system. However, there is a difference between the system states defined in the old state transition table and the new state transition table. This is because the combination of jobs in each computer changes due to the addition of jobs and computers. Therefore, in order for the user to change the system, the state transition table must be defined again in consideration of all the system states, and it must be verified whether or not this definition is correct. As described above, the state transition table exponentially increases according to the number of computers and the number of tasks to be added, and thus the work involved in this redefinition is not easy. Also, if the actual system configuration or event and the processing defined in the state transition table are different, the system will not operate normally.Therefore, when replacing the state transition table, the entire system must be stopped. Mandatory. This is also the case when deleting existing jobs or computers. Without stopping the entire running system, unnecessary jobs are removed from the system definition, and that much memory and disk space are used for other purposes. It cannot be reused.

Further, in the conventional system, the state transition table becomes large as the number of jobs increases, so it is difficult to increase the number of objects (businesses) to be managed independently. For this reason, managing execution in more detailed units, for example, in units of each program that constitutes a task, is itself extremely difficult to define. Therefore, in the conventional method, it is not possible to construct an operation management mechanism that enables the takeover from the middle of the work.

This invention does not increase the complexity of system definition due to an increase in the number of jobs / the number of computers, and makes it easy to add / delete jobs or add / delete computers without stopping the entire system. Aim to get.

[0016]

In a data processing apparatus according to the present invention, a plurality of job control tables defining job control information for each job, and a job control table search for searching a job control table corresponding to the job Means, and control information acquisition means for reading control information corresponding to an event from the retrieved job control table,
And a job control means for executing a job according to this control information.

Further, a plurality of job executing means, an abnormality detecting means for detecting an abnormality in the job executing means and generating an event, an execution priority table defining the priority of the job executing means for executing the job, and this priority. A job execution control means for causing another job execution means to execute the job executed by the abnormal job execution means based on the order table.

Further, it is characterized in that it has an execution priority table defined for each job.

In addition, it is characterized by having a job control table adding means for adding a job control table.

Further, it is characterized by having a job control table deleting means for deleting the job control table.

In addition, it is characterized in that it has a job execution means addition means for adding information of a new job execution means to the execution priority table.

Further, the present invention is characterized by further comprising a job executing means deleting means for deleting information on a part of the job executing means from the execution priority table.

Further, a program control table provided for each program corresponding to a plurality of programs constituting a job and defining processing when an event occurs during program execution, and a plurality of job executions for executing the job Means, a job execution abnormality detecting means for detecting an abnormality in the execution of the job and generating an event, and a program executed in the job showing the abnormality is identified based on the execution state, and the identified program is Job execution control means for causing another job execution means to execute the processing defined in the corresponding program control table.

Further, a plurality of job executing means for executing a job composed of a plurality of programs whose execution order is predetermined, and an abnormality detecting means for detecting an abnormality in the job executing means and generating a job executing means abnormal event Means, a program control table that is provided for each program and that defines an abnormal process when a job execution means abnormal event occurs, and an execution state recording means that records the execution state of the program executed by the job execution means,
Execution job execution based on this execution priority table that defines which job execution means will preferentially execute the job executed on the abnormal job execution means when an event occurs, and this execution priority table Job executing means determining means for determining means, interrupting program specifying means for specifying an interrupting program executed by the abnormal job executing means when an abnormal event occurs, and program control corresponding to the interrupting program An abnormal process defined in the table and a job execution control unit that causes the inherited job execution unit to execute the interrupted program.

[0025]

In the data processing apparatus configured as described above, a plurality of job control tables that define job control information for each job, job control table search means that searches the job control table corresponding to the job, The job control definition includes a control information acquisition unit that reads control information corresponding to an event from the retrieved job control table and a job control unit that executes the job according to the control information. Is sufficient and works to reduce the definition of job control.

Further, a plurality of job executing means, an abnormality detecting means for detecting an abnormality in the job executing means and generating an event, an execution priority table defining the priority of the job executing means for executing the job, and this priority. The job control table defines job control information for each job by including a job execution control unit that causes another job execution unit to execute the job that was being executed by the abnormal job execution unit based on the order table. However, since the priority order of the job executing means to be executed is defined in the execution priority table, the work of defining the job control information and the execution order table works independently.

Further, by having the execution priority table defined for each job, it works to define the priority of the job executing means for executing the job for each job.

In addition, by having a job control table adding means for adding a job control table, job control information is defined for each job in the job control table, and the job is added using the job control table adding means. It works when you do not change other job control tables.

Further, by providing a job control table deleting means for deleting the job control table, the control information of the job is defined for each job in the job control table, and the job is deleted by using the job control table deleting means. In addition, it works so as not to change other job control tables.

In addition, by having the job execution means addition means for adding the information of the new job execution means to the execution priority table, when the new job execution means is added to the execution priority table, the job control table is added. Work so as not to change.

Further, by having a job executing means deleting means for deleting information on a part of the job executing means from the execution priority table, when a part of the job executing means is deleted from the execution priority table, the job is deleted. It works so as not to change the control table.

Further, a program control table provided for each program corresponding to a plurality of programs constituting a job and defining processing when an event occurs during program execution, and a plurality of job executions for executing the job Means, a job execution abnormality detecting means for detecting an abnormality in the execution of the job and generating an event, and a program executed in the job showing the abnormality is identified based on the execution state, and the identified program is By having the job execution control means for causing the other job execution means to execute the processing defined in the corresponding program control table, the program control table for one program is the program control table for the other program. If the job ends abnormally, the recorded real To identify the program that was running the state serves to execute this program.

Further, a plurality of job executing means for executing a job composed of a plurality of programs whose execution order is predetermined, and an abnormality detecting means for detecting an abnormality of the job executing means and generating a job executing means abnormal event Means, a program control table that is provided for each program and that defines an abnormal process when a job execution means abnormal event occurs, and an execution state recording means that records the execution state of the program executed by the job execution means,
Execution job execution based on this execution priority table that defines which job execution means will preferentially execute the job executed on the abnormal job execution means when an event occurs, and this execution priority table Job executing means determining means for determining means, interrupting program specifying means for specifying an interrupting program executed by the abnormal job executing means when an abnormal event occurs, and program control corresponding to the interrupting program By having the abnormal processing defined in the table and the interrupted program, and the job execution control means for causing the inherited job execution means to execute, the program control table for one program is related to another program. Abnormality in one job execution unit independent of the program control table When it occurs, the interrupted program being executed is identified from the recorded execution state, and the takeover job executing means, which is another job executing means, operates to execute the interrupted program based on the priority table. .

[0034]

【Example】

Example 1. A first embodiment according to the present invention will be described for a hot standby type dual system. FIG. 1 is an overall configuration diagram of a computer system according to the present invention. In FIG. 1, 10a and 10b are computer A and computer B respectively (hereinafter, when the description is made without distinguishing either one,
(Computer 10), 11 is a processor in the computer 10 that performs data processing, 12 is a memory that is a storage device that is connected to the processor 11 and stores data, programs, etc. 13 is two computers 10 and A shared storage device, which is connected by a cable 14 and stores data and the like of these computers 10, 15 is two computers 1
A LAN cable connecting 0s and 16 is management information which is data for managing the operation of this computer system.

The computer system according to the first embodiment is
It uses a hot standby dual system. The hot standby type dual system consists of an operating system and a standby system.
It is composed of one computer, and if the operating system stops for some reason, the standby system takes over the work. For example, computer A
When the computer 10a operates in the operating system and the computer B10b is in the standby system, the computer A10a normally executes all the tasks, and the computer B10b stands by while monitoring the operation of the computer A10a. If the computer A10a fails for some reason, the computer B10b detects this and takes over all the work performed by the computer A10a and executes it instead of the computer A10a. Here, the job refers to a job including a plurality of related program groups or a single program. Further, since the work is handed over according to the management information 16, the management information 16 is changed when the management information 16 is changed so that the management information of both the computers 10a and 10b is the same.
The management information 16 of b is changed. A management program that manages the computer system and performs the above-described business takeover processing, management information change processing, and the like runs on each computer 10.

FIG. 2 is a block diagram of the management information 16 of the dual hot standby system that executes two tasks, and has a function equivalent to that shown in the state transition table of FIG. In the computer system shown in FIG. 1, these pieces of information are defined in each of the computers 10 constituting the system (the content of the information is the same among all the computers 10).

First, the components of the management information 16 shown in FIG. 2 will be described. Reference numeral 101 is an operating state table that stores the state of the work being executed by the computer 10.
Reference numeral 11 is a priority table (which is an execution priority table) that defines which computer 1 is to execute the job 1 preferentially.
12 is the same as 111, the priority table for the job 2, 12
Reference numeral 1 is a state transition table for work 1, and 122 is a state transition table for work 2.

The state transition tables 121 and 122 shown in this figure
Defines the current business status in rows and events in columns, and defines the processing and new business status under each condition. The number following the arrow in the table is the number indicating the new business state, and the diagonal line means that the combination does not occur. (In FIG. 2, processing when an event occurs is omitted.). The priority tables 111 and 112 are arranged in descending order of priority, and in FIG. 2, the computer 1 is designated in descending order of priority from the left. In this case, since it is a dual hot standby system, the contents of the priority table of both jobs are the same, and the computer A1 is in descending order of priority.
0a and computer B10b. Therefore, in the normal state, the computer A10a is an operating system and executes the tasks 1 and 2. When the computer A 10a stops, the computer B 10b takes over the tasks 1 and 2. In addition, in order to manage the work executed in each computer 10 in the system,
It has a driving state table 101. This table has entries for the number of computers (2 in this case), and manages whether the computers A10a and B10b are operating and which work is being executed, respectively. At the time of business takeover, each computer entry in the operating state table 101 is sequentially referred to according to the contents of each priority table. For this reason, pointers are provided from the respective entries of the priority order tables 111 and 112 to the respective computer entries of the operating status table 101.

Further, the state transition tables 121 and 122 store the address where the command and the like are stored for each event for each state. Taking the state transition table 121 of FIG. 2 as an example, the number of events is 3 (starting, abnormal termination, restoration completion) and the number of states is 3 (stopping, executing, restoring), so the state transition table 121 is 3 X3 (number of events x number of states)
With the address of. Then, in the memory pointed to by this address, the process (command) when an event occurs in a certain state and the number of the state to be moved to next are specified. For example, if an activation event occurs while the state is stopped, the memory indicated by 121a in FIG. 2 is checked and the address stored therein is read. next,
The command and the next state number stored in the memory indicated by this address are read. This command is made up of a program name and an argument, and this command is executed (here, the startup program for job 1 is executed). Then, the next state number (2) described above is stored as the current state (here, it is changed from "stopping" to "running").

In this example, the recovery operation performed by the computer B 10b when the computer A 10a is stopped for some reason will be described. The following series of processing is performed,
The tasks 1 and 2 executed in the operating system (computer A10a) are taken over by the standby system (computer B10b).

1. First, the computer B 10b detects the stop of the computer A 10a, and the operating state table 101
The entry of the computer A10a is in operation → stopped. In a state where the computer A10a is operating normally, the management program in each computer 10 communicates with the computer 10 or the state of work periodically or as needed through the LAN cable of FIG. For example, computer A10a
When a new job occurs, the management program stores the occurrence of the job in the state management data in the memory 12 of the computer A10a. At the same time, the management program of the computer B10b notifies (broadcasts) the occurrence of work to other computers (here, the computer B10b). The computer B10b receiving this notification rewrites the management information 16 in the same manner as the computer A10a based on the notification. In addition, the computer A 10a and the computer B 10b periodically send an operation confirmation packet to notify the management program of the other party that the own computer 10 is operating. Therefore, for example, when the computer A10a is stopped, the above-mentioned operation confirmation packet is not transmitted, and therefore the computer B1
0b detects that the computer A10a has stopped when the above packet is not received for a certain time or longer.

2. Next, from the operation state table 101,
It is understood that the tasks 1 and 2 have been executed, and the abnormal stop of the tasks 1 and 2 is detected. The management program in the computer B10b is configured so that when the computer A10a stops, the management program
Examine the work that a was performing. The operating state table 10 of FIG. 2 stored in the memory 12 of the computer B10b
It can be seen from 1 that the computer A10a was executing the work 1 and the work 2, and therefore an abnormal end event is generated to notify that the work 1 and the work 2 have abnormally ended.

3. Upon receiving the above abnormal end event, the processing defined in the abnormal end event of the state transition table in each work is performed. Usually, this process is often the process of restarting the business after executing the recovery process.
Here, the state transition tables 121 and 122 are defined so that the jobs are restarted when the recovery is completed for both the jobs 1 and 2. (At this time, as a means for retrieving the state transition table corresponding to the business, for example, for the state transition table existing for each business, a table (not shown) of the addresses on the memory 12 in which these are stored is used. There is a method of providing the data on the memory 12 indicated by the addresses stored in this table as a target state transition table.
Consider a case where there are tasks 1 to 3 (each task is given an identification number or the like). In the table, an identification number corresponding to a job and an address storing each state transition table are stored as a set (in this case, a total of 3 sets of data are stored). When searching the state transition table corresponding to the business, the address stored in the state transition table is read from the table based on the identification number corresponding to the business, and the data on the memory 12 indicated by this address is stored in the state transition table. Specify as. ) Perform the following processing for each of tasks 1 and 2. (In the case of this example, the same operation is performed for both business 1 and business 2. Therefore, they are collectively shown below.) (A) Computer 10 to execute (computer 10 to take over business)
To determine the priority, the priority table 111 (or 112) is referred to. Although the computer A10a is defined at the head of the priority table 111 (or 112), the operating state table 101 shows that the computer A10a is stopped and the computer A10a is excluded. (B) The computer B10b is defined in the next entry, and this computer is in operation. Therefore, the computer B10b is determined as the computer 10 that executes the process. (C) The restoration process of the work 1 (or the work 2) is performed by the computer B10b. (D) When the recovery process is completed, a recovery completion event occurs, and the computer B 10b restarts the business program according to the state transition table 121 (or 122).

Further, in the first embodiment, the computer 10 is considered as the job executing means, and the processing as described above is performed between the different computers 10. However, the job executing means is regarded as one processor, and the inside of one computer is considered. When there are a plurality of processors (for example, the processors are connected via a bus), the same function can be realized between the different processors.

As can be seen by comparing FIG. 17 of the conventional system and FIG. 2 of the present embodiment, it is possible to construct an equivalent system that takes the state shown in FIG. 16 with less definition information than the conventional system. . Further, since each definition information is simple, it is possible to perform more detailed management.

Example 2. A second embodiment according to the present invention will be described for a load sharing type dual system. For the sake of simplicity, the case where two computers as shown in FIG. 1 are used is shown. In the load share type dual system, both of the two computers 10 are operating systems, and when one computer 10 is stopped for some reason, the other computer 10 is operated.
To take over the business. FIG. 4 is a configuration diagram of the management information 16 of the dual load sharing system that executes two tasks. FIG.
In the above, the same reference numerals represent the same or corresponding parts. This management information 16 is defined for each computer 10 that constitutes the system (the content of the information is the same for all computers 10).

Since the constituent elements of the management information 16 are the same as those in the first embodiment, only the differences will be described. While the priority table 111 of the job 1 gives the highest priority to the computer A10a, the priority table 111 of the job 2 has the same priority. The priority table 112 is different in that the computer B10b is designated as the highest priority. This means that in the initial state, the work 1 is first executed by the computer A 10a and the work 2 is executed by the computer B 10b. Therefore, when both computers 10 are operating normally, they are both executing work and operating, and are not divided into an operating system and a standby system like a hot standby system.

The operation of the computer system according to the second embodiment is basically the same as that of the first embodiment. However, Example 1
When the computer A10a is stopped for some reason as compared with the case described in the operation described above, the computer A10a is executing only the work 1, and the computer B10b which inherits the execution of the work 1 is different. 2 is executed, while the restoration process of the succeeding job 1 is performed.

As can be seen by comparing FIG. 19 of the conventional system and FIG. 3 of the present embodiment, it is necessary to construct an equivalent system that takes the state shown in FIGS. 16 and 18 with less definition information than the conventional system. You can Further, as can be seen by comparing FIG. 2 and FIG. 3, in the load share type and the hot standby type,
Only the definition contents of the priority order table are different, and operation management can be performed in any type of dual system by the same management method. In other words, according to this management method, in the case of the hot standby type, all the operations can be regarded as a load share type special type in which all the operations are concentrated on one computer (= all the operations have the same priority setting). .

Example 3. A case where the present invention is applied to a multiple system will be described as a third embodiment. According to the present invention, the multiple system is an extension of the dual system, the management structure and the like do not change, and the definition information simply increases. FIG. 4 is a configuration diagram of the management information 16 of the triple load sharing system that executes three tasks. 4, the same reference numerals as those in FIG. 2 represent the same or corresponding parts.
However, the running state of the computer defined in the operating state table 101 is increasing as the number of computers 10 is expanded to three, and the priority tables 111 and 112 are both defined for the same reason. The number of computers being used is increasing. (The execution state table 101 and the priority order table 111 and 112 are not fixed length data, but change according to the number of computers in the system). Reference numeral 113 is a priority order table for work 3, which is the same as 111 or the like.
Reference numeral 3 is a state transition table for work 3, which is similar to 121 or the like.

In the third embodiment, a recovery operation will be described when the computer A 10a is stopped for some reason. Basically, it is the same as the case of the dual system except that the number of computers C10c and work 3 is increased.

1. Computer B10b and Computer C10c
Detects the stop of the computer A10a, and the entry of the computer A10a in the operating state table 101 is operating →
It will be stopped.

2. From the operating status table 101, it is found that the work A is being executed on the computer A 10a, and an abnormal end event for the work 1 occurs.

3. Computer B10b and Computer C10c
Performs the process for the abnormal end event defined in the state transition table 101 of task 1. Usually, after the recovery process is executed, the business is often restarted. Therefore, in this example, the state transition table 1 is set so that the restart is performed after the recovery process is completed.
01 is defined. The processing for the task 1 is performed as follows. (A) To determine the computer to be executed (the computer that inherits the task 1), the priority table 101 of the task 1 is referred to. Although the computer A10a is defined at the head of the priority table 111, the computer A10a is
Is stopped, and the same judgment is made for the next entry in the next operating state table. (B) The computer B is added to the next entry in the priority table 111.
10b is defined and is in operation, the computer that executes the process is determined to be computer B10b. (C) The computer B 10b performs the recovery process of the task 1, and the process associated with the abnormal end event that has occurred in the computer C 10c ends. (D) When the restoration process is completed, a restoration completion event occurs, and according to the state transition table 321, the computer B10b
Then, the program for job 1 is restarted.

Next, the operation when a plurality of computers are stopped in the multi-system computer system will be described. here,
In the example of FIG. 4, the case where the computer A 10a and the computer C 10c are stopped in this order will be described. However, computer A
Since the stop of 10a is as described above, it is omitted.
Computer C10 after completion of taking over the business of computer A10a
The stop of c will be described below. (Current status is Computer A
10a is stopped, the computer B10b is executing the tasks 1 and 2, and the computer C10c is executing the task 3. )

1. Calculator C1 by Calculator B10b
The stop of 0c is detected, and the entry of the computer C10c in the operating state table 101 is in operation → stop.

2. From the operating state table 101, it is found that the business C3 is being executed on the computer C10c, and an abnormal end event for the business C3 occurs.

3. The computer B10b processes the abnormal end event defined in the state transition table 123 of the job 3. Usually, after the recovery process is executed, the business is often restarted. Therefore, in this example, the state transition table 123 is defined so that the restart is performed after the recovery process is completed. The computer B 10b performs the following processing for the task 3. (A) The priority table 113 is referred to in order to determine the computer to be executed (the computer that takes over the task 3). Although the computer C10c is defined at the top of the priority table 113,
From the operating state table 101, it is found that the computer C10c is stopped, and the next entry in the operating state table 113 is checked. (B) The computer A is added to the next entry in the priority table 313.
10a is defined, but similarly, the computer A 10a is stopped, so the next entry is checked. (C) The computer B10b is defined in the next entry, and it can be seen from the operating state table that the computer is in operation, and the computer that executes the processing of task 3 is computer B10b.
To decide. (D) By the above processing, the computer B10b recognizes that the computer B10b itself executes the task 3, and therefore the computer B10b performs the restoration process of the task 3. (E) When the restoration process is completed, a restoration completion event occurs, and according to the state transition table 123, the computer B10b
Then, the program for job 3 is restarted.

As can be seen by comparing FIGS. 3 and 4, the difference between the dual system and the multiple system in this system is only the following two points, and both systems can be constructed by the same management information structure. .

1. As the definition information, there are many state transition tables 123 and priority table 113 for business 3. However, this is due to the increase in the number of operations and is not an essential difference between dual / multiple systems.

2. In the priority table of each job, the number of entries for the computer C10c is increasing. This is the point that will change due to the increase in computers. Also, in this example, the load-share type multi-system system was described.
The same applies to the hot standby multi-system system. In the case of the hot standby type, it can be seen as a load share type special form in which M computers (M <N) that do not execute work in the standard state are prepared for backup. Only the definition of the priority table for each job is different, and the management / operation can be performed by the exact same mechanism as described here.

Process Flow of Management Program The process of the management program shown in the first to third embodiments is shown in a flowchart as shown in FIGS. less than,
FIG. 5 showing the operation of the management program at the time of starting the computer 10 and the event processing when the computer 10 is performing normal processing will be described separately.

When the computer 10 starts up When the computer 10 starts up, the management program is loaded and processing starts. First, starting from step S0 in FIG. 5, it is determined in step S1 whether there is another computer 10 operating. Whether or not the computer 10 is operating can be checked by communicating with another computer 10 via the LAN cable 15 of FIG. Here, if it is determined that there is no other operating computer 10, the process proceeds to step S3, and if it is determined that there is another operating computer 10, the process proceeds to the next step S2.

Step S2 is a process of acquiring the management information 16 from another computer 10 that is already operating. If there is another computer 10 that is operating, some business or the like has already occurred in that computer 10, and the operating status and business status of the other computer 10 are acquired as the management information 16. This is because it is necessary to share the same information. Next, it moves to step S3.

Step S3 is a process of performing the processes of steps S4 to S6 for all n jobs executed on the computer system. Therefore, when the processes of steps S4 to S6 have been executed for n jobs, the process proceeds to step S7 of FIG. 6, and otherwise, the process proceeds to step S4 for the remaining jobs.

In step S4, the job i is another computer 10
To determine if it is running. Job i is any one of jobs 1 to n. In steps S4 to S6, the jobs 1 to n are sequentially executed as the job i. The determination in step S4 is based on the management information 16 acquired in step S2.
It is judged by looking at the data. Here, when it is determined that the work i is being executed by another computer 10, it is not necessary to execute the work i by the own computer 10, and the process returns to step S3. If it is determined that it has not been executed, the process proceeds to step S5.

In step S5, the priority table of the job i (priority table 1 for the job 2 according to the example of FIG. 4).
It is checked whether the head of 12) is the own computer 10.
If it is the own computer 10, the process proceeds to step S6. If it is not the own computer 10, the process returns to step S3.

Step S6 is a process for generating a start instruction event for the job i. When the activation instruction event occurs, the activation of the work i is started. When this process ends, the process moves to step S3 and the process for the next work starts.

Event Processing Next, the event processing when the start-up processing of the computer 10 is completed and the processing shifts to normal processing will be described with reference to FIG. During normal processing, the management program runs on another computer 10.
While communicating with each other regularly or as needed, the management information 16 is matched with each other, and the other computers 10 are monitored for normal operation. During this time, the computer 10 operates other programs in parallel. The event process is to perform a recovery process or the like in response to an event that occurs when a job is abnormally terminated or a stoppage of another computer 10 is detected while performing such normal process. Processing.

First, when the starting process is completed in step S3 of FIG. 5, the process proceeds to step S7 of FIG. In step S7, it is determined whether there is an event that has not been processed. Usually, an event has not occurred in many cases, and the process returns to step S7 again, and this processing is repeated and waits until an event occurs. When an event occurs, the process branches to steps S8, S11 and S14. The branch destination is determined by the event that has occurred. When an event that occurs when the other computer 10 is stopped is detected, the process proceeds to step S8. Any work (work i)
When the abnormal end event is detected, step S1
Move to 1. If a recovery completion event that occurs when the job i has completed recovery or a start instruction event for the job i is detected, the process proceeds to step S14.

(Stop Other Computer) Step S8 is a process for checking the work being executed by the other stopped computer 10. Since the work being executed by the other computer 10 is stored in the operating state table 101, the computer 10 which has stopped
Check all the jobs that were executed in step S9
Move on to. Step S9 is a process of individually generating an abnormal end event for all the jobs checked in step S8. When this process ends, the process returns to step S7.

(Business Abnormal Termination) In step S11, it is checked whether or not all the computers 10 having a higher priority than the own computer 10 are stopped, and it is determined whether or not the business i which has abnormally finished is restored by the own computer 10. Processing. The own computer 10 identifies the computer 10 in which the abnormally terminated job i should be executed by the priority table for the job i in the management information 16 (for example, the priority table 11 for the job 2 in FIG. 4).
2) Check and decide. If all the computers 10 having a higher priority than the own computer 10 are stopped (determined by referring to the operating state table 101), it is determined that the work i must be executed by the own computer 10. Then, the process proceeds to the recovery process of the next step S12, and otherwise returns to step S7.

Step S12 is a process for recovering the job i that ended abnormally. In this process, the business i is restored and the computer 10 is connected through the LAN cable 15.
Notifies that job i has started recovery. Further, the management information 16 in the own computer 10 is also rewritten during the recovery of the job i. When this process ends, the process moves to the next step S13.

Step S13 is a process for generating a recovery completion event after the completion of the recovery process started in step S12. When this process ends, step S7
Return to

(Business startup) Step S14 is a process for responding to the business recovery completion event generated in step S13 or the business startup event generated in step S15 of FIG. In step S14, similarly to step S11, it is determined whether or not all the computers 10 having higher priorities (for the target job i) than the own computer 10 are stopped. If it is determined that they are all stopped (YES), the process moves to the next step S15 and the work i is started. If not (NO), it is determined that the task i is not started in the own computer 10 and the step S7 is performed.
Return to

In step S15, the task i
Is a process for activating. Here, the operation state table 101 of the management information 16 of the self is started at the same time when the work i is started
Is rewritten to "execution of job i", and all the other computers 10 are notified that the job i has started up through the LAN cable 15. Other computer 10 that received this notification
Rewrites its own management information 16 to "(in the computer 10 that is the notification source) is executing the task i". After the end of step S15, the process returns to step S7 and stands by for the next event. Example 4. The addition of new work according to the present invention will be described below with reference to FIG. FIG. 7 shows another structure of the management information 16 shown in FIG. 1, which replaces the management information described in Embodiments 1 to 3 with reference to FIGS. 7, the same reference numerals as those in FIG. 4 represent the same or corresponding parts. Reference numeral 130 in FIG. 7 is a header part when searching the information of the computer 10, and the next state management data 131.1
A computer status header having 32 head addresses and connected by a pointer, 131 is connected by a pointer to the computer status header 130, and has the address of the next status management data and the head address of the business status of the computer A10a. Status management data, 132 also has the same information as 131, computer B
This is the state management data stored for 10b. 1
31 to 132 have a list structure, and new state management data can be easily added to the list structure. 141 and 142 are definition information headers (for example, 151.
152) and the next business status data (for example, 142) address, and the business status data is connected to each by a pointer. In the business status data, the business execution status (for example, stopped, in progress, in recovery, etc.) is stored. 143-1 and 141/1
This is a business status header that stores the same information as 42 for computer B10b. These business status data 141 to
144 is also divided for each computer 10 to form a list structure. Reference numeral 150 denotes a header part when searching for the definition information headers 151 to 154 of the business, and the definition information header 1 of the business
Business headers having head addresses of 51 to 154 and connected by a pointer, 151 to 154 are addresses of the next definition information header (for example, 152), addresses of a priority table (for example, 111), State transition table (for example, 1
21) and the like, and is a definition information header connected by a pointer to form a list structure. In FIG. 7, the arrow indicates that the pointer is connected, and the destination pointed by the arrow is the connection destination of the pointer.

Next, the operation of adding work will be described. The addition of a new job according to the present invention requires only the addition of a job priority table and a state transition table, and does not require any change to the existing part. Therefore, new work can be taken in without stopping the system during operation. For example, jobs 1 to 3 have already been executed on the computer 10,
A case where the work 4 is added here will be described. In FIG. 7, the data (1
14, 124, 144, 154) are not stored as the management information 16.

As described below, the processing is performed, and the new work is added without stopping the existing work. Currently 3 jobs
(Business 1, 2, 3) are executed by the computers A10a, B10b, respectively, and a new business 4 is added and this is executed by the computer B10b. This is a business setting,
It consists of two phases of business execution, and loads the setting information and executes the business according to the information. Each process is performed as follows, and the new job 4 is executed.

[Business settings]:

1. A memory area for the definition information header 154 is reserved and added to the work list. To add to the work list, the address indicating the next definition information header in the definition information header 153 of the work 3 is rewritten to the address of the newly secured memory area (that is, the pointer 160).
Setting).

2. The memory areas 124 and 114 for the state transition table and the priority order table of the task 4 are secured, and the memory area 12 for the state transition table and the priority order table is defined from the definition information header 154.
The pointers 164 and 163 to 4,114 are set.

3. The state transition table and the priority order table for job 4 are read into the prepared memory areas 124 and 114.
Now, the preparation for executing the task 4 is completed. The state transition table and the priority order table read here have a structure as shown in the state transition table 123 of FIG. 4, etc., and are defined by the user according to the work content. As shown in the state transition table 123 of FIG. 4 and the like, it has a very simple structure, and since the information to be defined only needs to define at least the information regarding the applicable business, the number of information to be defined can be small. Therefore, the definition information can be easily grasped. For example, when looking at the state transition table 123 of FIG. 4 and the like, the definitions for nine states (the actual operation is defined by four, and "/" is N
(ULL and the like are defined). This can be said for the work of rewriting the priority table when adding the computer 10 which will be described later. For example, the priority tables 111 to 114 are the priority table 10 of FIG.
When it has the same structure as 1, the number of pieces of information to be redefined is the same as the number of computers 10 and is small.

[Business execution]: 1. A memory area for the business status data 144 is secured, and a pointer 162 to the definition information header 154 of the business 4
Set.

2. By referring to the priority table 114 pointed from the setting information header 154, the computer B
It is decided to execute the task 4 in 10b. (The method of this determination has been described above with reference to FIG. 6.)

3. The business status data 144 is stored in the computer B10.
It is added after the business status data 143 of b, that is, the pointer 161 is set. Further, the addition of the job 4 is transmitted to each computer 10 in the system via the LAN cable, and the construction of the management information 16 as described above is also performed in another computer 10.

4. The computer B 10b activates the program according to the state transition table 124, and changes the business status in the business status data 144 from “stopped” to “execution in progress”. Also,
The computer B10b notifies each computer 10 in the system of the change in the business status.

As shown in this example, by managing information with an appropriate data structure, definition information can be added without affecting existing parts. In addition, each business is independent of each other, and execution of new business does not affect execution of other business. For example, when adding work 4,
Since the state transition tables 121 to 123 and the priority order tables 111 to 113 of other tasks (tasks 1 to 3) are not changed, the operations of tasks 1 to 3 operating based on these state transition tables 121 to 123 and the like. There is no need to stop. Further, since the state transition tables 121 to 123 of the tasks 1 to 3 have not been changed, it is verified based on these state transition tables 121 to 123 whether the tasks 1 to 3 normally operate in all states. No need. Therefore, it is possible to add a new job while the existing job is being executed as in the present embodiment.

Example 5. The deletion of work according to the present invention will be described with reference to FIG. To delete a job according to the present invention, it is sufficient to delete the priority order table and the state transition table for that job, and there is no need to change the definition part other than the job. Therefore, the work can be deleted from the system without stopping the entire system during operation. less than,
An example of dynamically deleting a business will be described.
FIG. 8 shows the management information 16 according to the fifth embodiment. 8, the same reference numerals as those in FIG. 7 represent the same or corresponding parts.

Currently, four jobs (jobs 1, 2, 3, 4) are being executed on the computers A10a and B10b, respectively.
Shall be stopped and removed from the system. When deleting a running job, delete the definition information after stopping the job. The processing is performed as follows.

1. In the computer B10b, the state transition table 12
4, the stop process of the work 4 is executed. At this time, the work 4 is set in the priority table 144 so that none of the computers 10 takes over. This setting is also made in the other computer 10, and the other computer 10 does not take over the work based on the priority table of the work 4.

2. The business status data 144 is stored in the computer B10.
Delete from the list of b. This deletion is performed by changing the address of the business status data next to the business status data 143 to "there is no next business status data and the end of the list (for example, NUL.
L) ”(that is, pointer 16
1 is cut).

3. The memory area for the business status data 144 is released. 4. The definition information header 154 is deleted from the work list (the pointer 160 is disconnected). 5. Definition information header 154, status transition table 12 for job 4
4. Free the memory area for the priority table 114.

As shown in this example, by managing the information with an appropriate data structure, the definition information can be deleted without affecting other parts. In addition, each work is independent of each other, and the suspension of one work does not affect the execution of the other work. Therefore, it is possible to delete a specific job without stopping the entire system as in the present embodiment.

Example 6. The addition of the computer 10 according to the present invention will be described with reference to FIG. In the addition of the computer 10 according to the present invention, the computer 1 that constitutes the system
Since the information about 0 is collected in the priority table of each work, the change of the priority due to the addition of the computer 10 does not affect the definition content of the state transition table. Further, the priority table needs to be referred to only when the system is started and when the work is handed over, and a new computer can be added without stopping the system during operation. The sixth embodiment in which a new computer is dynamically loaded will be described below. FIG. 9 shows the management information 16 according to the sixth embodiment. 9, the same reference numerals as those in FIG. 7 represent the same or corresponding parts. Reference numeral 133 is state management data for the computer C10c, which has the same structure as 131 and 132. 111a, 112a, and 113a are newly set priority tables and have the same structure as 111 and the like.

Here, an operation example in the case of adding the new computer C10c will be described. Currently 3 jobs (Job 1,
2 and 3) are respectively executed by the computers A10a and B10b, a computer C10c is newly added, and the task 2 that is being executed by the computer A10a is executed. The addition processing of the new computer 10 consists of two phases of business setting and business execution, and processes changing of the priority order tables 111 to 114 and business migration. The processing is performed as follows, and a new computer C10c is added to the system,
The job 2 is moved from the computer A10a to the computer C10c.

[Business settings]:

1. State management data 13 of computer C10c
Reserve memory area for 3 and add to list. To add the state management data to the list (131/132),
From the end of the list (business status data 132) to computer C1
This is done by setting a pointer to the state management data 133 of 0c. At this time, the pointer 16 to the business status data in the status management data 133 of the computer C10c
6 is NULL.

2. Since the entries of the computer C10c increase in each of the priority tables 111, 112, 113, the memory areas 111a, 112a, 113a are reallocated. At this time, the pointers from the definition information headers 151, 152, 153 to the old priority table 111, 112, 113 are cut off, and the new priority table 111a, 112a, 11
The pointer 167 to 3a is reset.

3. Reassigned priority tables 111a and 11a to which the priorities of the tasks including the new computer C10c have been reassigned
Read in the area of 2a and 113a. Here, since the job 2 is executed by the new computer C10c, the priority table 112a of the job 2 has the highest priority of the computer C10c, and is in the order of the computer A10a and the computer B10b (computer A10a and computer B10b). The order of B10b may be changed). Management information 1 for the above business settings 1 to 3
6 is operated by another computer 10 on the computer system.
The same is done in.

[Business execution]:

1. In order to move the work 2, the work 2 currently being executed by the computer A10a is stopped. Stop processing is executed in accordance with the state transition table 122, and the business status in the business status data 142 is changed from “in execution” to “stopped”. In addition, the computer A10a changes this business state by L
The other computers 10b and 10c are notified through the AN cable 15. (Here, in order to take over the work, it is only temporarily stopped, and by the subsequent processing, it is being executed again under the computer C10c.)

2. By referring to the priority order table 112a pointed to by the definition information header 152 of the job 2, it is decided to execute the job 2 on the computer C10c.

3. The business status data 142 is stored in the computer A10.
It is deleted from the list of a and added to the list of the computer C10c. To delete the business status data (142) from the list of the computer A10a, the pointer is followed from the status management data 131 of the computer A10a, and the business status data (141) immediately before the business management data (142) to be deleted.
To search. Then, the business data (14
1) The pointer 168 to the next business data is set to the business management data next to the business management data (142) to be deleted (in this example, since this business management data does not exist, NULL indicating the end of the list is set). Setting)). Business status data (142) is calculated by computer C
To add to the list of 10c, the pointer is traced from the status management data 133 of the computer C10c to search the end of the list of business status data, and from this last business status data to the business status data (142) to be added. This is done by setting the pointer (166). However, in this example, since the business status data of the computer C10c does not yet exist, the pointer 166 is directly set from the status management data 133 to the business status data 142. The operation for the management information 16 of this job execution 3 is performed by the computer system L.
It is performed on all computers 10 connected to the AN cable 15.

4. State transition table 1 on computer C10c
22 to start the program, the business status data 14
Change the job status in 2 from "Stopping" to "Running".
Further, the computer C10c notifies the other computer 10 of the change in the business state through the LAN cable.

As shown in this example, by managing information with an appropriate data structure, it is possible to add computer status data and change the priority table associated with the addition of computers without affecting other parts. It will be possible. Also, the priority table is
Since it is referenced only when the system is started and when the work is handed over, changing the priority table does not affect the work execution. Therefore, it is possible to add a new computer without stopping the entire system as in the present embodiment.

Embodiment 7 FIG. FIG. 1 shows the operation of disconnecting (deleting) the computer 10 from the system according to the present invention.
It will be described using 0. In the deletion of the computer 10 according to the present invention, the information concerning the computer 10 constituting the system is the priority table of each job (for example, 111a to 113a).
Since it is summarized in a), the change of the priority order of the computer does not affect the definition contents of the state transition table. Further, the priority order table (for example, 111a to 113a) needs to be referred to only when the system is started and when the work is taken over. Therefore, the existing computer 10 can be deleted from the system without stopping the running system.

FIG. 10 shows the structure of the management information 16 described in the subsequent operation. 10, the same reference numerals as those in FIG. 9 represent the same or corresponding parts.

A case where the computer C10c is deleted from the system will be described. Currently, three tasks (tasks 1, 2, and 3) are being executed on computers A10a, B10b, and C10c, respectively, and task 3 that was being executed on computer C10c is computer A.
Shall be taken over.

1. Since the number of entries of the computer C10c decreases in each of the priority order tables 111 to 113, the memory area 1
Reassign 11a, 112a, 113a. At this time, as in the case of the sixth embodiment, the definition information header 1
The pointers of the priority order table are reset from 51 to 153.

2. Information of the priority order of each job from which the computer C has been deleted is reassigned to each priority table 751, 752,
The area 753 is read, and pointers are set to the priority table 111a, 112a, 113a of each newly allocated computer. Here, in order to execute the work 3 on the computer A10a (or, if the computer C10c is deleted according to the priority order of the original priority table 113, it is automatically executed on the computer A10a having the next priority order). Newly assigned state management data 111a.1
In 12a and 113a, the priority order corresponding to each work is set. (Here, in the priority table 113a of the job 3, the computer A10a is set as the highest rank, and the computer B1 is next.
It is set so that 0b comes. ) In addition, by the notification from the LAN cable 15, the computer C
The management information 16 is similarly changed not only in 10c but also in other computers 10.

3. In order to move the work 3 to the computer A 10a, the work 3 currently being executed by the computer C 10c is stopped. The computer C10c is the state transition table 123 of the task 3
Execution of stop processing according to
The job status in 42 is changed from "execution" to "stopped". In addition, the computer C10c changes this business state by L
Notify each computer 10 through the AN cable. Receiving this notification, each computer 10 changes the work status data of its own work 3 from "running" to "stopped". (The suspension of business 3 here is to allow business 3 to succeed.
It is only temporarily stopped, and it is being executed again under the computer A10a by the subsequent processing).

4. By referring to the priority order table 113a (the computer C is deleted) pointed from the definition information header 153 of the job 3, it is determined that the computer A executes the job.

5. The job status data 143 of job 3 is deleted from the list of the computer C10c and added to the list of the computer A. The deletion and the addition are performed in the same manner as the job execution 3 described in the sixth embodiment (the pointer 169 is cut and the pointer 166 is newly set).

6. The program is started according to the state transition table 123, and the business status in the business status data 143 is changed from "stopped" to "execution in progress". Also, computer A10a
Notifies the other computer 10 of this change in business status,
The other computer 10 receiving this changes the management information 16 in the same manner.

7. Business status data 14 of computer C10c
3 is deleted from the list of business status data (pointer 16
6 is disconnected), and the memory area for the business status data 143 of the computer C is released. The processings 4 to 6 described above are also performed on the management information 16 in another computer 10.

As shown in this example, by managing the information with an appropriate data structure, the state management data (for example, 133) is deleted and the computer 10 is added without affecting other parts. Priority table (111-113 or 11
1a to 113a) can be changed. Further, the priority order table (111 to 113 or 111a to 113a) is referenced only when the system is started up and when the work is taken over, so that changing the priority order does not affect the execution of the work.
Therefore, as in the seventh embodiment, when the existing computer 10 is executing a task, only that task needs to be temporarily stopped (in order to be handed over to another computer 10).
It is not necessary to stop the entire system, and the existing computer 10 can be deleted while other tasks are being executed.

Example 8. Normally, a business is composed of a plurality of programs which are related to each other, but in the method of managing a business as a minimum unit, each business (program) creator must manage the program within the business. In the eighth embodiment, as an extension of the first to seventh embodiments,
Hierarchical task definition and execution status management
By expanding to ≧ 2), it is possible to manage at the level of the program that constitutes the business or at a more detailed level. Further, since the status of the work can be managed at the program level, it becomes possible to continue the work interrupted by the stop of the computer 10.

FIG. 11 shows the management information 1 according to the eighth embodiment.
6 shows the structure of 6. 11, the same reference numerals as those in FIG. 7 or 9 represent the same or corresponding parts. In the eighth embodiment, in addition to the state transition tables 121 to 124 for the work, the operation for the state change is defined for each program that constitutes the work. First, the components will be described. The management information 16 is roughly divided into two pieces: management information on the execution state and system definition information. The management information on the execution state is the same as in the fourth to seventh embodiments. The status of each computer 10 is independent of the status management data 131-133.
, And these are connected by a pointer from the computer header 130 to form a list structure. In addition, each computer 10
Each of the jobs executed in 1. has job status data 141 to 144 for managing the execution status of each job. These business status data 141 to 144 are status management data 131 to 133 of the computer 10 in which the business is executed.
Are linked by pointers to form a list structure. Further, since each business status data 141 to 144 needs the definition information of each business, it has pointers to the definition information headers 151 to 154, respectively. However, since the figure becomes complicated, in FIG.
1-114 to state management data 131 to 13 of each computer
The pointer to 3 is omitted.

The definition information part is almost the same as that shown in FIG. 7 etc., and the operation definition (program definition information 171-1
75) is the only difference. Also, with this,
The process definition (command address) at the time of state change, which has been included in the state transition tables 121 to 124, is unnecessary. The processing at the time of event occurrence is performed according to the above-described operation definition (program definition information 171 to 175).

The definition information is the state transition tables 121 to 121 of each job.
124, priority table 111-124, each program definition 171-178, and definition information header 151 of each job
˜154. The state transition tables 121 to 124 store the state transition of each work, and the priority order tables 111 to 114 store the priority order of the computer 10 in each work. In addition, since it is necessary to refer to the computer status at the time of business takeover, from each entry of the priority table 111 to 114,
According to the definition contents, the status management data 13 of each computer
1 to 133 (omitted in FIG. 11).

The definition information headers 151 to 154 pair the state transition tables 121 to 124 and the priority order tables 111 to 114 for each business (for example, the definition information header 15 of the business 1).
1 stores the state transition table 121 of the work 1 and the priority order table 111 of the work 1 as a pair and stores them. There are as many jobs as there are jobs, and a state transition table (for example, 121) and a priority table (for example, for each job)
111), a pointer to the program definition (for example, 171) of the program to be executed first in the business, and each definition information header 151 to 154 has a list structure in which the business header 150 is the head. Program definition 171
.About.178 have the following data, respectively. Here, since the information at each program level is sufficient, the definition information is generally summarized as follows. In the following sentences, the following labels (program, abort, rerun, nex) are used to indicate each component in the operation definition information of the program.
t, parallel, status).

[Program]: Program name Specify the program to be executed. [abort]: Abnormality processing program Specify the program to be executed when the program terminates abnormally due to the stop of the computer that was executing or the program itself. [rerun]: Restart when the program is finished. If you do not need to restart the program with either "Always restart", "Restart only when abnormal termination" or "Do not restart", the program execution is completed. To do. [next]: Program to start next Specify the program to be executed next after the execution of the program is completed. In terms of data structure, it is a pointer to a structure that defines another program. (For example, pointers 181, 182, 184 in FIG. 11) [parallel]: Program to be operated in parallel Designate programs to be executed simultaneously. The data structure is a pointer to a structure that defines another program.
(For example, the pointer 183 in FIG. 11). The destination of the parallel pointer becomes a subtree, and the completion of the subtree and the completion of the program specified by program are synchronized. [status]: Execution status of program Execution status of program (Before starting / Running / Ended / Abended)
Data area for recording.

The above-mentioned data is summarized as program definition information 171 to 178, which are connected by pointers (the above-mentioned next pointer and parallel pointer) in the order of execution to form a tree structure.

The execution order of the programs for the job 1 executed by the computer A10a in FIG. 11 when the system is managed by this data structure is shown below. Also, at the time of execution / end of each program, the status is recorded in the status of the program definition information 171 to 178.

[During normal processing]

1. The program definition information 171 directly pointed to by the job 1 is searched. Then, the program defined in the program of the program definition information 171 is executed. At this time, the LAN cable 15 indicates that the execution of the program has started on the other computer 10 as well.
Tell through.

2. Parall in the program definition information 171
Since the el pointer is not specified, the completion of the program being executed by the above operation 1 is waited for, and the processing target is moved to the program definition information 172 specified by the next pointer 181 of the program definition information 171. Here, the program defined in the program definition information 172 is executed.

3. On the other hand, since the program definition information 174 is designated in the program definition information 172 using the parallel pointer 183, the program definition information 1
The program defined in program 74 is executed in parallel (with the program executed in the above 2). At this time, the other computers 10 are also notified that the execution of the program has started. (A) The program definition information 174 contains a next pointer 1
Since the program definition information 175 is designated as 84, the completion of the program being executed in the program definition information 174 is waited for, and the processing of the program definition information 175 is executed (again, the execution of the program defined in program is executed. Etc.). (B) Wait for completion of the program executed by the program definition information 174, and complete the subtree starting from the program definition information 174. Further, the completion of the program execution is also notified to the other computers 10 through the LAN cable 15. 4. Program definition information 174 and program definition information 1
Wait for the completion of the partial tree of 75, and then program definition information 172
Processing of the program definition information 173 designated by the next pointer 182 is executed, and the program defined by program is executed. The completion of the program and the start of the next program are notified to the other computer 10 as described above. 5. The completion of the program executed in the above 4 is waited, the completion is notified to the other computer 10, and the execution of the program related to the job 1 is ended.

On the other hand, if an error occurs during job execution, the following processing is performed. The job execution abnormality includes the execution abnormality of the job itself and the abnormality of the computer 10. First, a case will be described in which the job 10 is abnormally terminated due to a failure in the job itself while operating normally.

[At abnormal end]:

1. When a job terminates abnormally, the program definition information (for example, 174) of the program being executed at that time is referred to, and this program definition information is aborted.
Executes the error handling program defined by the label of t. Also, abnormal termination is stored in status. The start of execution of the abnormal processing program is notified to the other computer 10 as in the case of the normal processing. At this time, as a case where the business ends abnormally, for example, an abnormality is detected by an interrupt process that occurs when a program created by the user runs out of control and tries to execute something that is not in the instruction set of the processor as an instruction There are cases where the work cannot be continued. As a method of identifying which program of the job has terminated abnormally, there is a method of determining that the program definition information whose data stored in status is "in execution" is the program that aborted due to an error. .

2. Wait for the completion of the error handling program,
The other computer 10 is notified of the end of the abnormality processing program.

3. Write end to status and perform processing according to the next program definition information (indicated by next).

Next, the case where the job is abnormally terminated because the computer 10 is stopped will be described. When the computer 10 that is executing work is stopped, the stopped computer 10
It is determined that the work being executed in step 3 was abnormally terminated.
This judgment is made by another computer 10 and the processing is carried out in the same manner as shown in the first embodiment. However, in the eighth embodiment, the processing at the time of abnormal termination is the above [at the time of abnormal termination]: 1 to
The difference is that it is performed by another computer 10 as in the case of 3. Processing such as determination of the computer 10 that executes a job is performed in the same manner. In the above explanation, after the abnormal processing is executed after the occurrence of the abnormality, the program shown in the next next is to be moved.However, the program shown in the next next is the one after the current processing is completely finished. If it is premised that the program is interrupted, the program interrupted by the error is executed in the error process, and then the next process is performed. However, since there are various possible program anomalies, the user is free to control during the anomaly process whether to execute the interrupted program or to take a temporary measure and move to the next (next) program. be able to.

As described above, since the relationship between the programs constituting the business is reflected in the management information 16 and the execution status of each program is managed, the computer 10 which takes over the business uses this information as a basis. You can decide where to resume execution (= program). Therefore, it is not necessary to restart the work from the first program, and the program can be executed halfway while inheriting the state (program unit) during execution. Therefore, it is possible to reduce the time required for the handover process.

In this example, the end tree is always synchronized with the end of the tree, but a flag may be provided for the end tree so that the end synchronization is not set. Further, the abnormality processing program is set to one program, but the abnormality processing program may be selected depending on the end code of the business program. For example, when the exit code of the program being executed is "-1", the abnormality processing program specified in [abort1] is executed, and when it is a value other than "-1", it is specified in [abort2]. For example, another abnormality processing program may be executed.

Example 9. Management information 1 described in the eighth embodiment
Processing for adding a job in No. 6 will be described as Embodiment 9 with reference to FIG. According to the N-stage (N ≧ 2) state management shown in the eighth embodiment, it is possible to manage at the level of the program that constitutes the work, and it is possible to continue from the middle of the work. Also,
In this configuration, the state management in each business unit shown in Embodiments 1 to 7 is also left, and each business is managed independently of each other. Therefore, even in this case, it is possible to add a new task during system operation.

FIG. 12 illustrates the structure of the management information 16 according to the ninth embodiment. 12, the same reference numerals as those in FIG. 7 or FIG. 11 represent the same or corresponding parts. Further, in the ninth embodiment, it is assumed that the new job 4 is added to the computer C10c, and the job 4 is composed of the programs shown in Table 1 and has the priority shown in Table 2.

Table 1 shows the programs defined in the program definition information constituting the work 4.

[Table 1]

Table 2 describes the priority table of job 4.

[Table 2]

The processing procedure of addition and activation of the work 4 added to the computer C10c of FIG. 12 is shown below.
The addition of work is basically the same as that shown in the fourth embodiment. However, in the ninth embodiment, the program definition information 17
1 to 180 are newly added, the program definition information 171 to 180 is read from the shared storage device 13 of FIG.
The difference is that 78 to 180 are set in each computer 10 and that the business is started according to this program definition information. This process will be described below.

1. A memory area for the definition information header 154 of the job 4 is allocated.

2. A memory area for the priority order table 114 is allocated, the priority order of Table 2 is read from the shared storage device 13 of FIG. 1, and the pointer 163 of the priority order table 114 is set from the definition information header 154 of the job 4.

3. A memory area for the state transition table 124 is allocated, and the state transition table 124 is stored in the shared storage device 1 of FIG.
Then, the pointer 164 from the definition information header 154 of the job 4 to the state transition table 124 is set.

4. From the contents of Table 1 stored in the shared storage device 13 and the like, it can be seen that there are three programs forming the job 4, and the program definition information 178 to 180
Allocate a memory area for.

5. The definition information of Table 1 is read into each of the memory areas (178 to 180). As a result, the next pointer 186 of the program definition information 178 points to the program definition information 179, and the next pointer 186 of the program definition information 178
The parallel pointer 187 points to the program definition information 180.

6. Since the first program to be executed from the contents read in 5 above (Table 1) is program1, the definition information header 154 of the task 1 is the program definition information 1
Point 78 (pointer 185).

7. The definition information header 154 of the job 4 is added to the list of definition information headers (definition information header 153
To the definition information header 154 is set).

8. The computer 10 that executes the job 4 is determined by referring to the priority table 114 (which is the computer C10c).

9. A data area for the business status data 144 is allocated, and a pointer 162 to the definition information header 154 of the business 4 is set.

10. The business status data 144 is stored in the computer C1.
0c added to the list of status management data (work status data 143 of job 3 to job status data 144 of job 4)
Pointer 161 to () is set.

11. When executing a task, the program group is sequentially activated according to the setting of the program definition information 178.

As described above, the management information 16 is changed in each computer 10 connected to the LAN cable 15 (however, the portion related to the execution of the added work is the computer 10 instructed to execute (this execution). In Example 9, computer C1
0c) only).

As shown in this example, even when the execution management is performed at the program level, it is possible to add the definition information without affecting the existing part by managing the information with an appropriate data structure. Therefore, it is possible to add a new job while the existing job is being executed.

Example 10. A process of deleting a job in the management information 16 described in the eighth embodiment will be described as the tenth embodiment with reference to FIG. N stages shown in Example 8 (N ≧
According to the state management of 2), it is possible to manage at the level of the program that constitutes the business, and it is possible to continue even after the business.
Further, in this configuration, the state management in each business unit shown in the first to seventh embodiments is left, and each business is managed independently of each other. So even in this case,
Jobs can be deleted without stopping the entire system.

FIG. 13 illustrates the structure of the management information 16 according to the tenth embodiment. In FIG. 13, the same reference numerals as those in FIG. 7 or 11 represent the same or corresponding parts.
Further, in the tenth embodiment, of the two jobs (jobs 3 and 4) executed by the computer C10c, the job 4 is deleted. In FIG. 13, the priority order table 11
1-114 to state management data 131 to 13 of each computer
The pointer to 3 is omitted.

With respect to the work 4 added to the computer B10b shown in FIG. 13, the procedure for stopping and deleting the work to be deleted is shown below. The deletion of work is basically the same as that shown in the fifth embodiment. However, in the tenth embodiment, as in the ninth embodiment, since the program definition information 171 to 180 is newly added, the point that the program definition information 178 to 180 related to the task to be deleted is deleted from each computer 10, and The difference is that the business is stopped according to the program definition information. The deletion processing procedure for the job 4 deleted from the computer C10c in FIG. 13 is shown below.

1. From the definition information header of job 4, job 4
By referring to the respective program definition information 178 and 180 constituting the above, each running program is stopped. This time,
The priority table 114 of each computer 10 is set so that none of the computers 10 will take over the task 4.

2. The business status data 144 is stored in the computer C10.
It is deleted from the work list of c (pointer 162 is cut), and its memory area 144 is released.

3. Each program definition information 178/180
Free the memory area of. 4. The memory areas of the priority order table 114 and the state transition table 124 are released. 5. The definition information header 154 of the job 4 is deleted from the job list (the pointer 160 is disconnected), and the memory area of this definition information header 154 is released. As described above, the management information 16 is changed in each computer 10 connected to the LAN cable 15.

As shown in this example, even when the execution management is performed at the program level, the definition information can be deleted without affecting the parts other than the deleted part by managing the information with an appropriate data structure. Yes, there is no effect on the deletion work other than the deletion work, and the specific work can be deleted while the other works are being executed.

Example 11. A process of adding the computer 10 in the management information 16 described in the eighth embodiment will be described as the eleventh embodiment with reference to FIG. According to the state management of N stages (N ≧ 2) shown in the eighth embodiment, it is possible to manage at the level of the program that constitutes the work, and it is possible to continue from the middle of the work. Further, in this configuration, the state management in each business unit shown in the first to seventh embodiments is left, and each business is managed independently of each other. Therefore, even in this case, it is possible to add the new computer 10 during system operation.

FIG. 14 illustrates the structure of the management information 16 according to the eleventh embodiment. 14, the same reference numerals as those in FIG. 7 or 11 represent the same or corresponding parts.
Further, in this embodiment, it is assumed that the computer D10d is added and the job 4 is moved from the computer C10c to the added computer D10d. In addition, in FIG. 14, the priority table 1
11 to 114 state management data 131 to 1 of each computer
The pointer to 33 is omitted.

With respect to the processing for adding the computer C10c in FIG. 14, the processing procedure of the computer addition and job transfer processing will be shown below. The process of adding a computer and moving the work is basically the same as that shown in the sixth embodiment. However, in the twelfth embodiment, as in the eighth to tenth embodiments, the program definition information 171 to 180 is newly added, so that the suspension of the work is different in that the definition contents of the program definition information 178 to 180 are followed.

1. State management data 13 of computer D10D
Allocate a memory area for 4.

2. State management data 13 of computer D10D
4 is added to the list of state management data (131 to 133) (pointer 189 from the state management data 133 of the computer C10c to the state management data 134 of the computer D10d).
Set).

[0167] 3. Old priority table for each task (computer D10
111.112.113.114 is released, and areas 111a.112a.113a.114a for the new priority table (including computer D10d) are allocated.

4. The new priority table (including the computer D10d) is read into 111a, 112a, 113a, 114a.

5. Definition information headers 151 to 15 for each job
4 sets pointers 190 to the corresponding priority tables 111 to 114, respectively.

6. In order to move the work 4 to the computer D10d, the work 4 being executed by the computer C10c is once stopped. (This is a temporary stop. Job 4 is restarted on the computer D10d in the subsequent processing.)
Further, this stoppage is transmitted to another computer 10 through the LAN cable 15.

7. Business status data 14 of computer C10c
From the list of 3.144, the status management data 14 of the job 4
4 is deleted (the pointer from the business status data 143 to the business status data 144 is disconnected).

8. According to the new priority table 114 of the work 4, the computer 10 that restarts the work 4 is determined. (As described above, in this example, the computer D10d is restarted.)

9. The job 4 is restarted on the computer D10d, and the status management data 114 of the job 4 is added to the list of job status data of the computer D10d (in this case, the pointer 188 of the job status data 144 from the status management data 134 of the computer D10d). Set).

As shown in this example, even when the execution management is performed at the program level, by managing the information with an appropriate data structure, the priority order tables 111 to 114 can be changed without affecting the existing part. Therefore, the new computer 10 can be added while the existing business is being executed (except when the existing business is handed over to the new computer 10). Further, even if the existing work is handed over to the new computer, it is not necessary to stop the other work.

Example 12. A process of deleting the computer 10 in the management information 16 described in the eighth embodiment will be described as the twelfth embodiment with reference to FIG. According to the N-stage (N ≧ 2) state management shown in the eighth embodiment, it is possible to manage at the level of the program that constitutes the work, and it is possible to continue from the middle of the work. Further, in this configuration, the state management in each business unit shown in the first to seventh embodiments is left, and each business is managed independently of each other. Therefore, even in this case, it is possible to delete the computer during system operation.

FIG. 15 illustrates the structure of the management information 16 according to the twelfth embodiment. 15, the same reference numerals as those in FIG. 7 or FIG. 11 represent the same or corresponding parts.
In this embodiment, the computer B10b is deleted and the computer B1 is deleted.
It is assumed that the task 2 executed in 0b moves to the computer A10a. In FIG. 15, the pointer from the priority list to the status management data of each computer is omitted.

With respect to the processing for adding the computer C10c in FIG. 14, the processing procedure of the computer deletion and job migration processing will be shown below. The processing of deleting a computer and migrating work is basically the same as that shown in the seventh embodiment. However, in the tenth embodiment, as in the eighth to eleventh embodiments, since the program definition information 171 to 180 is newly added, the processing such as the stop due to the movement of the work is performed by the program definition information 178 to 181.
The difference is that it follows the definition content of 0.

1. Old priority list for each task (computer B
10b) (including 10b) 111 to 114 are released, and memory areas for the new priority lists 111a, 112a, 113a, 114a (not including the computer B10b) are allocated. )

2. The new priority list (not including the computer B10b) is replaced with the new priority list 111a / 112a.
・ Read to 113a ・ 114a.

3. Definition information header 131 to 13 of each job
3 to the corresponding priority table 111a, 112a, 113
The pointer 190 to the a 114a is reset.

4. In order to delete the computer B10b, it is necessary to let the other computer 10 take over the task 2 executed by the computer B10b. Since the job 2 is moved to the computer A10a and executed, the job 2 is stopped once. (This is a temporary stop. Job 2 is restarted on the computer A10a in the subsequent processing.)

5. Business status data 13 of computer B10b
Status management data 132 of the job 2 from the list of 1 to 133
Is deleted (the pointer 191 is set from the state management data 131 of the computer A10a to the state management data 133 of the computer C10c).

6. According to the new priority list 112a of the work 2, the computer that takes over the work 2 is determined. (As described above, in this example, the computer A10a takes over.)

7. The job 2 is restarted on the computer A10a, and the status management data 142 of the job 2 is added to the list of the job status data 141 of the computer A10a (from the job status data 141 of the job 1 to the job status data 142 of the job 2).
Pointer 192 to) is set). 8. Calculator B10b
The memory area of the state management data 132 of is released.

As shown in this example, even when performing execution management at the program level, it is possible to delete a computer while executing an existing job (except when the computer to be deleted is executing the job). . Further, even when the computer to be deleted is executing the business, it is not necessary to stop the other business.

[0186]

Since the present invention is constructed as described above, it has the following effects.

A plurality of job control tables defining job control information for each job, job control table searching means for searching the job control table corresponding to the job, and control according to the event from the searched job control table Since the control information acquisition unit for reading out information and the job control unit for executing the job according to the control information are provided, it is sufficient to define the job control for one job, and the definition amount of the job control is reduced. Therefore, it is possible to easily define the control information of the job with respect to the system state.

Further, a plurality of job executing means, an abnormality detecting means for detecting an abnormality of the job executing means and generating an event, an execution priority table defining the priority of the job executing means for executing the job, and this priority The job control table defines job control information for each job by including a job execution control unit that causes another job execution unit to execute the job that was being executed by the abnormal job execution unit based on the order table. However, since the priority order of the job executing means to be executed is defined in the execution priority table, the control information of the job for the system state can be easily defined regardless of the number of job executing means on the system.

Further, by having the execution priority table defined for each job, the priority of the job executing means for executing the job is defined for each job. Therefore, the job executing means for executing the job for each job is defined. Can be controlled.

In addition, by having the job control table adding means for adding the job control table, the control information of the job is defined for each job in the job control table, and the job is added by using the job control table adding means. In case no other job control table is changed,
New jobs can be added without stopping the jobs that are operating according to other job control information.

Further, when the job control table deleting means for deleting the job control table is provided, the control information of the job is defined for each job in the job control table, and the job is deleted by using the job control table deleting means. Moreover, since the other job control tables are not changed, some jobs of the plurality of jobs can be deleted without stopping the jobs operating according to the other job control information.

In addition, by having the job execution means addition means for adding the information of the new job execution means to the execution priority table, when the new job execution means is added to the execution priority table, the job control table is added. Since it does not change, job execution means can be added without stopping the job.

Further, by having a job executing means deleting means for deleting information on a part of the job executing means from the execution priority table, when a part of the job executing means is deleted from the execution priority table, the job Since the control table is not changed, it is possible to delete a part of the job executing means without stopping the job being executed by the job executing means not to be deleted.

Further, a program control table provided for each program corresponding to a plurality of programs constituting a job and defining processing when an event occurs during program execution, and a plurality of job executions for executing the job Means, a job execution abnormality detecting means for detecting an abnormality in the execution of the job and generating an event, and a program executed in the job showing the abnormality is identified based on the execution state, and the identified program is By having the job execution control means for causing the other job execution means to execute the processing defined in the corresponding program control table, the program control table for one program is the program control table for the other program. If the job ends abnormally, the recorded real By executing this program by identifying the program that was running from the state, you can easily create a program control table, and if necessary, do not rerun the job from the beginning, Execution can be resumed from a program in the middle of.

Further, a plurality of job executing means for executing a job composed of a plurality of programs whose execution order is predetermined, and an abnormality detecting means for detecting an abnormality in the job executing means and generating an abnormal event in the job executing means. Means, a program control table that is provided for each program and that defines an abnormal process when a job execution means abnormal event occurs, and an execution state recording means that records the execution state of the program executed by the job execution means,
Execution job execution based on this execution priority table that defines which job execution means will preferentially execute the job executed on the abnormal job execution means when an event occurs, and this execution priority table Job executing means determining means for determining means, interrupting program specifying means for specifying an interrupting program executed by the abnormal job executing means when an abnormal event occurs, and program control corresponding to the interrupting program By having the abnormal processing defined in the table and the interrupted program, and the job execution control means for causing the inherited job execution means to execute, the program control table for one program is related to another program. Abnormality in one job execution unit independent of the program control table When it occurs, the interrupted program that was being executed is identified from the recorded execution status, and the takeover job execution means, which is another job execution means, executes the interrupted program based on the priority table. You can easily create a control table, and if necessary, you can restart the job from the interrupted program in the middle of the job based on the above record without re-executing the job from the beginning. The priority of the job executing means to be executed can be controlled.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a computer system for explaining a first embodiment of the present invention.

FIG. 2 is a configuration diagram of management information according to the first embodiment of the present invention.

FIG. 3 is a configuration diagram of management information according to a second embodiment of the present invention.

FIG. 4 is a configuration diagram of management information according to a third embodiment of the present invention.

FIG. 5 is a flow chart illustrating an operation at the time of starting a computer according to the third embodiment of the present invention.

FIG. 6 is a flowchart illustrating an operation when an event occurs according to the third embodiment of the present invention.

FIG. 7 is a configuration diagram of management information for explaining addition of work according to a fourth embodiment of the present invention.

FIG. 8 is a configuration diagram of management information for explaining business deletion according to the fifth embodiment of the present invention.

FIG. 9 is a configuration diagram of management information for explaining computer addition according to the sixth embodiment of the present invention.

FIG. 10 is a configuration diagram of management information for explaining computer deletion according to the seventh embodiment of the present invention.

FIG. 11 is a configuration diagram illustrating management information according to an eighth embodiment of the present invention.

FIG. 12 is a configuration diagram of management information for explaining addition of work according to a ninth embodiment of the present invention.

FIG. 13 is a configuration diagram of management information for explaining business deletion according to the tenth embodiment of the present invention.

FIG. 14 is a configuration diagram of management information for explaining computer addition according to the eleventh embodiment of the present invention.

FIG. 15 is a configuration diagram of management information for explaining computer deletion according to the twelfth embodiment of the present invention.

FIG. 16 is a state transition diagram illustrating a conventional computer system.

FIG. 17 is a configuration diagram of conventional management information.

FIG. 18 is a state transition diagram illustrating a conventional computer system.

FIG. 19 is a configuration diagram of conventional management information.

FIG. 20 is a diagram illustrating a system state in the conventional system when one computer is used and the number of tasks is three.

FIG. 21 is a diagram illustrating a system state in the conventional system when two computers are used and the number of tasks is three.

[Explanation of symbols]

10 computer, 11 processor, 12 memory,
13 shared storage device, 15 LAN cable, 1
01 operating state table, 111-114 priority table, 121-124 state transition table, 130 computer state header, 131-134 state management data, 14
1-144 business status data, 150 business header,
151-154 definition information header, 171-180
Program definition information

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Hiroyuki Yamanishi 325 Kamimachiya, Kamakura City Mitsubishi Electric Corp. Information Systems Works

Claims (9)

[Claims] 1. Job control information according to an event,
A plurality of job control tables stored for each job, a job control table search means for searching a job control table corresponding to an externally specified job from the plurality of job control tables, and a job control table search means for searching. The control information acquisition means for reading the control information of the job corresponding to the above event from the job control table and the job to be searched according to the event when the above event occurs and the job control table is instructed to search the job. And data processing including job control means for controlling the execution of the controlled job according to the control information acquired from the control information acquisition means, and job execution means for executing the job according to the control of the job control means. apparatus. 2. A plurality of the job executing means are provided, and when any one of the plurality of job executing means has an abnormality, the job executing means is identified as an abnormal job executing means and the abnormality has occurred. As the above-mentioned event for notifying that the job execution means has an abnormal event, an abnormal event detection means for generating an abnormal event, and when the abnormal event for the job execution means has occurred, the job being executed by the abnormal job execution means is executed by a plurality of other jobs The abnormal job executing means executes the job based on the execution priority table that defines which of the job executing means has priority to execute, and the execution priority information defined in the execution priority table. The job execution means that executes the job is selected as the selected job execution means, and 2. The data processing apparatus according to claim 1, further comprising: the job control unit, which commands execution of a job being executed by the abnormal job execution unit. 3. The data processing apparatus according to claim 2, wherein the contents of the execution priority table are defined for each job. 4. A storage device for storing the job control table, and a job control table adding means for storing the new job control table in an unused area in the storage device. 3. The data processing device according to any one of 3 above. 5. The job control table deleting means for deleting an arbitrary job control table from the plurality of job control tables stored in the storage device, according to any one of claims 2 to 4. The described data processing device. 6. The data processing apparatus according to claim 2, further comprising a job execution means addition means for adding information on a new job execution means to the execution priority table. 7. The data processing apparatus according to claim 2, further comprising a job executing means deleting means for deleting information on a part of the job executing means from the execution priority table. 8. A program control table that is provided corresponding to each of a plurality of programs that make up a job and that defines processing when an event occurs during program execution, and a plurality of program control tables that execute a job composed of a plurality of programs. Job execution means, execution state recording means for recording the execution state of the program of the job executed by the job execution means, and job execution abnormality detection means for detecting the execution abnormality of the job and generating the event And, when the event is detected, the program that was being executed by the job that showed an abnormality is specified based on the execution status, and the processing defined in the program control table corresponding to the specified program is executed. A data processing device, comprising: another job execution control unit that causes the job execution unit to execute the job execution unit. 9. A plurality of job executing means for respectively executing a job constituted by a plurality of programs whose execution order is predetermined, and the plurality of job executing means. When an abnormality occurs, this job executing means is identified as an abnormal job executing means, and an abnormality detecting means for generating a job executing means abnormal event is provided, and the job executing means abnormal event is provided in each of the plurality of programs. The program control table that defines the abnormal processing when it occurs, and the job that was being executed by the abnormal job execution means when the abnormal event of the job execution means occurred An execution priority table that defines whether the execution means should be executed preferentially, and this execution priority order Based on the table, the job execution means determining means for determining the job execution means for inheriting and executing the job executed on the abnormal job execution means as the inherited job execution means, and the job execution means for executing The execution status recording means for recording the execution status of the program of the job, and the interrupted program for executing the program executed by the abnormal job execution means based on the execution status when an abnormal event of the job execution means occurs. The interrupted program specifying means, the abnormal processing defined in the program control table corresponding to the interrupted program, and one of the interrupted program or the specific program in the abnormal job executing means, Job execution control means for causing the job execution means to execute Data processing device.