JP2003085021A - Batch processing system equipped with recovery/restart function, program for batch processing system equipped with recovery/restart function, and recording medium for recording program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system capable of restarting after half-finish of a batch job relative to a business database, while securing continuity of processing before and after the half-finish, as a batch processing system equipped with a recovery/restart function. SOLUTION: This system is constituted so that the progression state of the batch job relative to the business database DB by an application program AP, an input file FI and an output file FO is monitored and recorded by a check point table TB constituted as a database, and that the business database DB and the check point table TB are synchronously processed simultaneously. At the restart time after the half-finish, the job is started based on job progression management information recorded in the check point table TB.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer system, and more particularly, to a recovery system suitable for performing online processing and batch jobs in parallel on a common database.
A batch processing system having a restart function, a program for such a system, and a recording medium recording the program belong to the field of data processing technology using a computer.

[0002]

2. Description of the Related Art When a batch job terminates abnormally due to a failure, conventionally, depending on the cause, a database whose processing has stopped midway is replaced with a database backed up before the job starts, A measure has been taken such as starting a job again from the beginning or deleting the record of the input file up to the normally completed portion and editing the application program to restart from the continuation.

However, in the former case, when a plurality of jobs are concurrently performed on a common database, the data updated by a process other than the abnormally terminated process returns to the state before the update. , It becomes unknown to what part of the input file the update of the database by the latter process is reflected. Also, in the latter case, extremely troublesome work such as deleting the record of the input file and editing the application program is required.

In any case, other processing for updating or referencing the database must be stopped until the recovery of the database is completed. Especially when online processing is performed in parallel, this online processing is performed. The processing will be stopped for a long time, which will hinder the business.

Therefore, in order to solve this kind of problem,
The application program that executes the batch job for the database can recover the abnormal termination in advance.
While incorporating the logic for restart,
It has a management file that monitors the execution of batch jobs by this application program and records the records that have completed normally for each specified processing unit.When restarting after an abnormal end, the jobs are processed according to the information recorded in this management file. The method of starting is considered.

[0006]

However, the above method makes the application program complicated, significantly increases the development labor thereof, and lacks versatility. Particularly, the recovery restart is performed by using the existing application program. There is a drawback that it is difficult to add a function.

Further, according to the above method, even if the database synchronization processing and the recording in the management file are performed at the same time, only one of them is performed depending on the timing of failure occurrence and the contents of the failure. The system may be stopped, in which case the records in the management file and the contents of the database become inconsistent.

Therefore, according to the present invention, the recovery / restart after the abnormal termination due to the occurrence of the failure can always be correctly executed while ensuring the continuity of the database before and after the abnormal termination, and the recovery / restart is performed. A system whose functions are easy to realize and a program for the system,
Another object is to provide a recording medium having the program recorded therein.

[0009]

That is, the present invention is characterized by having the following constitution in order to solve the above problems.

First, the invention according to claim 1 of the present application relates to a batch processing system for executing a batch job using a business database by a batch job executing means, and more particularly to a batch processing system having a recovery / restart function. There is a progress management table configured as a database that records the progress management information of batch jobs,
A synchronization processing unit that records progress management information in the progress management table at a predetermined time during execution of the batch job, and simultaneously performs synchronization processing for the progress management table and synchronization processing for the business database, and a batch job At the time of start, based on the recorded contents of the progress management table, start state determination means for determining whether the current start is an initial start or a restart after a halfway end, and when the determination means determines restart, Based on the progress management information recorded in the progress management table, a start position locating means for locating the start position of the processing target data so that the batch job is restarted with the state when the synchronous processing is performed before the halfway end as the initial state. And having.

Here, the data to be processed includes an input file, an output file, a business database and the like.

The invention according to claim 2 of the present application is a system according to claim 1, particularly in a system in which a batch job execution means executes a batch job based on an input file to a business database. The processing means records the number of records input from the input file in the progress management table as progress management information at a predetermined interval based on the start of the job or the time of the previous synchronization processing, and the update status at that time The progress management table and the business database are synchronized with each other, the start position locating means positions the next record of the number of inputs recorded in the progress management table at the start position of the input file, and the batch job executing means executes the job. Position at the start position by the above positioning means in the input file at restart And executes a job from the record that has been.

According to a third aspect of the present invention, the synchronous processing means is also provided in the system according to the first aspect, particularly in the system in which the batch job execution means executes the batch job based on the input file to the business database. When the key break is detected every predetermined number of times in the input file, the number of records input from the input file and the contents of the record when the key break is detected in the input file are recorded in the progress management table as the progress management information. The synchronization processing of the business database is performed in the updated state by the record immediately before the key break, and the synchronization processing of the progress management table is performed in the updated state at the time of detecting the key break, and the start position locating means is recorded in the progress management table. Start record of the input file for the record next to the input count Positioning and batch job execution means, when a job restart, reads the record of the time of the key break detection from the progress management table,
It is characterized in that the record of the input file is read from the record positioned at the start position by the positioning means in the input file and the job is executed.

According to a fourth aspect of the present invention, in the system according to the first aspect, particularly in the system in which the batch job execution means executes a batch job based on the record of the business database, the synchronization processing means is The key value of the processed record of the business database is recorded in the management table as the progress management information, and the start position locating means processes the record of the key value next to the key value of the processed record recorded in the progress management table. The batch job execution means is positioned at the start position of the database, and when the job is restarted, the job is executed from the record positioned at the start position by the positioning means in the business database.

Further, in the invention described in claim 5, in the system according to any one of claims 1 to 4, particularly, in the system in which the batch job execution means processes the output file, the synchronous processing means is , The number of writing out data to the output file is recorded in the progress management table as the progress management information, and the start position locating means positions the data next to the number of writing out recorded in the progress management table at the start position of the output file. ,And,
The batch job executing means is characterized in that when the job is restarted, the job is executed from the data positioned at the start position by the positioning means in the output file.

According to the invention described in claims 1 to 5, while the batch job is being executed by the batch job executing means, the synchronous processing means performs the synchronous processing on the business database at a predetermined time. However, in the process of synchronizing the business database, progress management information such as the number of records input from an input file is recorded in a progress management table configured as a database, and the progress management table and the business It will be synchronized with the database at the same time. Therefore, the contents of the business database and the contents of the progress management table are always kept in a consistent state, and at the time of restarting after a halfway termination due to an error or the like, both of them are rolled back to the state of the same synchronization processing point. become.

Then, based on the recorded contents of the progress management table, when it is determined that the current start is a restart, the start position locating means records the time of the last synchronization processing recorded in the progress management table. Since the restart position of the data to be processed is located based on the progress management information, by starting the batch job from this restart position, the business database will not cause duplicate input of records or omission of records, Will maintain the continuity of data.

According to the second aspect of the present invention, the above operation is performed when a batch job based on an input file is performed with respect to the business database, and is based on when the job starts or when the previous synchronous processing is executed. According to the third aspect of the present invention, the above operation is also performed when a batch job based on the input file is performed on the business database. This is realized when the synchronization processing is performed every time a key break is detected in a record of the input file or when a plurality of key breaks are detected.

Further, according to the invention of claim 4, when a batch job based on the record of the business database is performed without using the input file, according to the invention of claim 5, the process for the output file is performed. In this case, the action of the invention of claim 1 is realized, and in any case,
The restart after the halfway is performed while maintaining the continuity of the data before and after the restart.

According to a sixth aspect of the present invention, in the system according to any one of the first to fifth aspects, the assistance for performing a predetermined auxiliary work during the execution of the job by the batch job executing means. The work management means is provided, and the progress management table is provided with an auxiliary work information recording field for recording information on the progress of the auxiliary work performed by the auxiliary work execution means. It is characterized in that auxiliary work information reading means for reading the information in the work information recording column as an initial value of the auxiliary work after restart is provided.

According to the present invention, during the execution of the job step, auxiliary work such as totalization processing of the number of updates of specific items in each record and total processing of items such as the amount of money is performed, and the result is obtained. The data is recorded in the auxiliary work information recording column provided in the progress management table, and at the time of executing the synchronous processing of the business database and the progress management table,
Information on this auxiliary work is also processed synchronously.

When the job is restarted after it is terminated, the information about this auxiliary work is read from the auxiliary work information recording column of the progress management table and set to the initial value of the work, so that the job is restarted. Before and after, the auxiliary work is executed while maintaining continuity.

Further, the invention according to claim 7 is to execute the system according to any one of claims 1 to 6, particularly to execute a batch job constituted by a plurality of job steps for each job step. In the system configured as described above, the progress management table is provided with a status recording column indicating the status of each job step,
Status recording means is provided for normalizing the status recording column of the job step when the job step ends normally, and when the batch job is restarted from the first job step, the job step is started at the start of each job step. Job status skip means is provided for reading the status in the status recording column and skipping the job step at the normal end.

According to the present invention, when a plurality of job steps forming a batch job are sequentially executed, for a job step that has ended normally, a status record column for the job step in the progress management table is
It will be recorded as a normal end in sequence. Then, when one of the job steps ends abruptly due to an abnormality or the like, and after recovery, when the batch job is restarted from the first job step, each job step is referenced by referring to the status record column of the progress management table. It is determined whether the job has ended normally, and therefore, by skipping the processing that ended normally, unnecessary execution of the job step is avoided.

On the other hand, the invention according to claim 8 is configured as a batch processing system having a recovery / restart function, particularly a business database used for batch jobs and a database for recording progress management information of the batch jobs. A program for a batch processing system provided with a progress management table, comprising: a batch job executing means for causing a computer to execute the batch job; and a predetermined time during execution of the batch job by the executing means. A synchronization processing unit for recording progress management information in the progress management table and simultaneously performing a synchronization process for the progress management table and a synchronization process for the business database, based on the recorded contents of the progress management table at the start of a batch job, This time the start is after the initial start or halfway A start state determining means for determining whether it is a start, and a state when a synchronization process is performed before the midway end based on the progress management information recorded in the progress management table when it is determined to be restarted by the determining means. It is characterized by functioning as a start position locating means for locating the restart position of the processing target data so that the batch job is restarted in the initial state.

The invention according to claim 9 is the program according to claim 8, particularly in a program used in a system in which a batch job execution means executes a batch job based on an input file to a business database. When the computer is made to function as the synchronization processing means, the number of records input from the input file is recorded as the progress management information in the progress management table at a predetermined interval based on the start of the job or the previous execution of the synchronization processing. At the same time, when it is made to function so as to perform the synchronous processing of the progress management table and the business database in the update state at that time, and when it functions as the start position locating means, the next record of the number of input records recorded in the progress management table is set. Function to position it at the start of the input file, One, when caused to function as a batch job execution means may be operable to execute a job from the record positioned in the starting position by the positioning means in the input file when the job restart.

The invention described in claim 10 is also the program according to claim 8, particularly in a program used in a system for executing a batch job based on an input file to a business database by a batch job executing means. When the computer is made to function as the synchronization processing means, when a key break is detected every predetermined number of times in the input file, when the number of records input from the input file as the progress management information in the progress management table and the key break of the input file is detected. And record the contents of the record of, and make the business database perform the synchronization processing in the update state of the record immediately before the key break and the synchronization processing of the progress management table in the update state at the time of detecting the key break, respectively. Functions as a start position positioning means When the job is restarted, the next record after the number of inputs recorded in the progress management table is positioned at the start position of the input file, and when it is used as a batch job execution means, the progress management is performed when the job is restarted. It is characterized in that the record at the time of detecting the key break is read from the table, and the record of the input file is read from the record positioned at the start position by the positioning means to execute the job.

According to the invention described in claim 11, in the program according to claim 8 as well, particularly in a program used in a system for executing a batch job based on a record of a business database by a batch job executing means, a computer is used. When functioning as the synchronization processing means, the progress management table is made to record the key value of the processed record of the business database as progress management information, and when it is made to function as the start position locating means, it is recorded in the progress management table. When the function that positions the record with the next key value of the processed record's key value at the start position of the business database and also functions as the batch job execution means, when the job is restarted, the above-mentioned positioning in the business database is performed. To start position by means Characterized in that to function so as to perform a job from the location Tagged records.

The invention according to claim 12 is the program according to any one of claims 8 to 11,
In particular, in a program used in a system in which processing is performed on an output file by a batch job execution means, when the computer is made to function as a synchronization processing means, the number of times data is written to the output file is recorded in the progress management table as progress management information. To function as a start position positioning means,
When functioning to position the next data of the number of writings recorded in the progress management table at the start position of the output file and to function as the batch job execution means, use the above positioning means in the output file when restarting the job. It is characterized in that it functions to execute a job from the data located at the start position.

In the invention described in claim 13, the program according to any one of claims 8 to 12, in particular, an auxiliary work information recording column for recording information on the progress of the auxiliary work is provided in the progress management table. In a program used for the system, a computer performs a predetermined auxiliary work during execution of a job by the batch job executing means, and records information on the progress of the auxiliary work in an auxiliary work information recording field of the progress management table. It is characterized by functioning as auxiliary work executing means and auxiliary work information reading means for reading the information in the auxiliary work information recording column as the initial value of the auxiliary work after the restart.

Further, in the invention described in claim 14, the program according to any one of claims 8 to 13, particularly, a batch job composed of a plurality of job steps is executed for each job step. In a program used in a system in which a status recording column indicating the status of each job step is provided in the progress management table, a computer is used to set the status of the job step in the progress management table when the job step ends normally. A status recording unit that normally ends the recording column, and when restarting a batch job from the first job step, read the status in the status recording column of the job step at the start of each job step, and Skip that job step Wherein the function as job step skip means.

According to the invention described in any one of claims 8 to 14, the system according to any one of claims 1 to 7 is configured by mounting the program on a computer. By operating this, the same operation as the system of claims 1 to 7 can be obtained.

The invention as set forth in claim 15 of the present application,
A main program for causing a computer to function as a batch job executing means, and the program according to any one of claims 8 to 14 to be called by the main program so that the computer functions as each means other than the batch job executing means. It is characterized in that it is configured with a subprogram to be performed.

According to the present invention, the program provided with the recovery / restart function is divided into the original main program for the batch job and the subprogram for the recovery / restart function.

Therefore, for example, an existing application program for a batch job can be easily modified into a program having a recovery / restart function by adding a subprogram, or without considering the recovery / restart function. It becomes possible to develop application programs for batch jobs. Further, when the batch job is composed of, for example, a job step that updates the database for each record of the input file, a job step that collectively updates multiple records at the time of a key break, or a job step that does not use the input file, etc. Even if the application program is different for each job step, it is possible to add the recovery / restart function to each job step by also using the subprogram.

Further, the invention according to claim 16 of the present application relates to a computer-readable recording medium in which the program according to any one of claims 8 to 15 is recorded. According to the invention, By operating a computer using the recording medium, the operation of the system according to any one of claims 1 to 7 and claim 15
The action for the program will be realized.

[0037]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below.

FIG. 1 shows a hardware configuration of a system according to this embodiment. This system is capable of executing batch jobs and online processing in parallel. A batch processing unit 11 and an online processing unit 12 are provided. As an external device, the batch processing unit 11 is connected to an operator terminal 21 for inputting various commands and monitoring the progress of jobs, and the online processing unit 12 includes a communication control device 22. A large number of online terminals (not shown) are connected to the communication control device 22 via communication lines.

Further, the host computer 10 has
Storage device 1 storing application programs and subprograms for batch jobs, respectively
3 and 14, a business database group DB to be managed by the system, and a checkpoint table TB configured as a database for recovery / restart of batch jobs, which are provided in the batch processing unit 1 described above.
Connected to 1. Further, the batch processing unit 11 is capable of reading data from the input file group FI and writing data to the output file group FO, and the batch processing unit 11 has a memory 15 required for a batch job. And a timer 16 are provided.

All or part of the above-mentioned business database group DB is also connected to the online processing section 12 and is also a target of online processing. Also, although not shown,
The online processing unit 12 is also provided with a necessary program recording device, a memory, and the like.

Next, the software configuration of this system, that is, the configuration of a program for operating the host computer 10, will be described. As shown in FIG. 2, the program of this system is composed of an application program group AP and a plurality of subprograms. And a subprogram group SP. In addition to these program groups AP and SP, an OS normally required in this type of system
(Operating system) and JCL (Job control language) are naturally installed.

Application program group AP
Is composed of a plurality of application programs that respectively execute a plurality of job steps that form a batch job, and executes processing for the business database group DB. When the system handles a plurality of batch jobs, the same application program group can be provided for each job.

On the other hand, the subprogram group SP includes a file open subprogram SP1, a file close subprogram SP2, an input file reading subprogram SP3, an output file writing subprogram SP4, and a recovery / restart subprogram S.
It is composed of P5 and a synchronization processing subprogram SP6.

Then, these subprograms SP1 to SP1
SP6 is called by the application program being executed as necessary, and has a file open subprogram SP1 and a file close subprogram SP2.
Are input file group FI and output file group F, respectively.
The input file reading subprogram SP3 executes the processing for the input file group FI, and the output file writing subprogram SP4 executes the processing for the output file group FO. Further, the recovery / restart subprogram SP5 executes processing for the checkpoint table TB, and the synchronous processing subprogram SP6 executes processing for the business database group DB and the checkpoint table TB.

Here, an example of the configuration of one database DB1 in the above-mentioned business database group DB will be explained. As shown in FIG. 3, the database DB1 of this example is a customer master used in a credit card management system, for example. The customer number is used as a key to record data such as each customer's name, home zip code, address, telephone number, work name, zip code, address, and telephone number.

A configuration example of one input file FI1 in the input file group FI for updating the contents of the business database DB1 will be explained. As shown in FIG. 4, the input file FI1 of this example is The records are arranged in order according to the record number, and data such as a customer number, a category indicating an update item, and the update content of the item is recorded in each record. Then, for example, the customer number “10010
When the work place of the customer of “1” is changed, the work name of the division “4”, the work postal code of the division “5”, the work address of the division “6”, the work telephone number of the division “7” A record for updating etc. is created.

Further, the above checkpoint table T
Explaining the configuration of B, as shown in FIG. 5, in this table TB, each job step name of a plurality of batch jobs is used as a key, and the number of records input from the input file is recorded for each job step. Input number record field,
A write block number recording field is provided to record the number of data blocks composed of multiple records to be written to the output file. Also, a plurality of totalization results are recorded to record various totalization results for job steps that perform totalization processing. A result recording column and a user-defined area used for recording other data required in each job step are provided, and a status recording column for recording the status of the step is further provided.

Here, this checkpoint table T
In B, a plurality of input case number recording fields and a plurality of write block number recording fields are provided for job steps that handle a plurality of input files and output files. Further, the user-defined area can be arbitrarily set for each job step.

Next, the specific operation of the batch job performed by the batch processing section 11 of the host computer 10 performed according to each program will be described with reference to the flowcharts shown in FIG. In the following description, the execution operation of the application program is called a "main routine" and the execution operation of the sub program is called a "subroutine".

Different application programs are used for each job step. Among them, as the first embodiment, first, the operation database DB1 shown in FIG. 3 is updated for each record of the input file FI1 shown in FIG. A case will be described below in which an application program of a type that performs the synchronization processing and automatically performs the synchronization processing at a predetermined interval after the start of work or the execution of the previous synchronization processing is used.

The job step by this application program is to call the recovery / restart subprogram SP5 in step S1 of the main routine whose flowchart is shown in FIG. 6 after initializing all variables and the like used in the following control. Be started.

The subroutine by the recovery / restart subprogram SP5 is executed according to the flowchart of FIG. 7. First, in step S21, it is determined whether the current processing operation is the processing at the start of the job step or the processing at the end of the job step. . This determination is made depending on whether or not the subprogram SP5 is called in step S1 of the main routine.
When it is called in step 1, it is determined that the current process is a start process, and otherwise, it is determined that the process is an end process.

Since the current process is the start process, the checkpoint table T is next determined in step S22.
Referring to B, it is determined in step S23 whether or not a record of the job step is created in this table TB.

Now, the job step is the job name JOB.
-Assuming ST3 of A, checkpoint table T
When referring to B, as shown in FIG. 5A, if the record of the job step ST3 has not been created yet, it is determined that the present start processing is initial start.

In this case, in step S24, the process shown in FIG.
The name “ST3” of the job step that started the processing this time is written in the job step name field indicated by the symbol a in FIG. In step S25, this job step ST
The status of No. 3 is set to "initial start", and this is recorded in the memory 15 shown in FIG. The status record column of job step ST3 of the checkpoint table TB is held blank.

On the other hand, as shown in FIG. 5B, when the record of the job step ST3 has already been created in the checkpoint table TB, the start of the current job step ST3 is restarted after the midway end. If there is, step S2 of the recovery / restart subroutine
In step 6, the data such as the totalization result indicated by the symbol b recorded in the user-defined area of the record is read onto the memory 15, and the status is set to "restart" in step S27, and this is recorded in the memory 15. . Even in this case, the status record column of the job step ST3 is held blank.

If the execution of this subroutine is not based on the call from step S1 of the main routine, that is, if it is the end processing of the job step, as will be described later, steps S21 to S2 are executed.
8. In this case, the status of the job step is set to "normal end" in step S28, the status is recorded in the memory 15, and the status recording column of the checkpoint table TB stores " Write "Normal end".

When the job step start process is completed as described above, the operation returns to the main routine, and in step S2 of the flowchart of FIG. 6, the status of the job step is read from the status record column of the checkpoint table TB, It is determined whether this is a "normal end". If it is not “normal end”, that is, if it is “initial start” or “restart”, then step S
In step 3, the file open subprogram SP1 is called.

The subroutine by this subprogram SP1 is executed as shown in FIG.
Then, it is determined whether the file to be opened first is an input file according to a predetermined order.

If it is an input file, it is opened in the input mode in step S32, and in step S33, the status of the job step recorded in the memory 15 is "initial start" or "restart". It is determined, and in the case of normal "initial start", step S3
At 9, it is determined whether or not there is a file to be opened next.

If the status recorded in the memory 15 is "restart", the restart position is positioned in step S34. That is, the opened file is, for example, the input file FI1 described above, and the file shown in FIG.
As shown in, when the record of the job step ST3 to be processed this time has already been created in the checkpoint table TB, the number of records input from the input file FI1 recorded in the record (in the case of the example in the figure,
100) is read and the next value is input file F
It is positioned as the restart position of I1, and this is recorded in the memory 15.

On the other hand, when the file to be opened this time is an output file, steps S31 to S35 are executed, and the status of the job step recorded in the memory 15 is "initial start" or "restart". To determine. And in the case of "initial start",
In step S36, the file is opened in the output mode, and in the case of "restart", the file is opened in the input / output mode in step S37. Here, the open in the output mode means a mode in which a new output file is created, and the open in the input / output mode means a mode in which an existing output file is opened.

In the case of "restart", after the output file is opened in the input / output mode as described above, the restart position is positioned in step S38. That is, the opened file is, for example, the output file FO1.
5B, the number of blocks written to the output file FO1 of the job step ST3 recorded in the checkpoint table TB shown in FIG. 5B (in the case of the example,
50) is read and the value next to that value is output to the file FO.
It is located at the restart position of 1 and is recorded in the memory 15.

In any case, after opening the first file, it is determined in step S39 whether there is a file to be opened next, and if there is a next file, the above steps S31 to S38. Are repeatedly executed, and the above-mentioned input file FI1 and output file FO1 are executed.
Open all the files necessary for the processing of this job step starting with.

When the file open subroutine is completed as described above, it is then determined in step S4 of the main routine in FIG. 6 whether the current start processing is "initial start" or "restart", and "initial start" is performed. In the case of “,” step S5
Then, the work area provided on the memory 15 for recording the status, recording the number of input / output cases, recording the totalization result, etc. is initialized. In the case of "restart", as described above, in the recovery / restart subroutine,
The tabulation result or the like is read from the checkpoint table TB to the memory 15, or the positioning processing of the input file and the output file is performed in the ill-open subroutine.

Then, in step S6, the input file reading subprogram SP3 is called. The subroutine executed by this program is executed according to the flowchart of FIG. 9. First, in step S41, it is determined whether or not the mode of the synchronization process for the database of the job step is the automatic synchronization mode based on the attribute of the application program. If it is the automatic synchronization mode, it is further determined in step S42 whether or not the input file for which the current record is to be read is the file to be synchronized, based on the attribute of the file. To do.

That is, since the input file reading subprogram SP3 is also used for a plurality of job steps having different synchronous processing modes for the business database, and is also used for reading an input file that is not the target of the synchronous processing, The determination in steps S41 and S42 is performed.

As described above, the application program for executing the job step this time updates the business database DB1 record by record, and automatically performs the synchronization process of the record.
Since the input file FI1 used in this processing is a file for synchronization processing, step S43 is executed next,
The synchronous processing subprogram SP6 is called. The subroutine of the synchronization processing executed by this synchronization processing subprogram SP6 will be described in detail later.

Thereafter, in this input file reading subroutine, the first record of the input file FI1 is read in step S44. In that case, the first record means that the execution start of this job step is "initial start".
In the case of, the record is the first record of the file, and in the case of "restart", it is the record positioned in step S34 of the file open subroutine of FIG. Then, after reading the record, in step S45,
Count up one entry. In that case, in the case of "restart", the count is incremented by 1 with respect to the number of inputs read from the checkpoint table TB.

Next, in step S7 of the main routine,
It is determined whether reading of all records of the input file FI1 has been completed. This determination is made when step S44 of the input file reading subroutine is executed.
It is determined whether or not a record is actually read, and if there is no record to be read, it is determined that the reading is completed.

Since the reading of the input file FI1 is not completed at the beginning of the job step, step S8 is executed next, and predetermined processing such as predetermined counting processing is executed together with the predetermined step. , Output file writing subprogram SP4 in step S9
To output the output data created based on the input record or the like to the output file via the external device.

The subroutine by the output file writing subprogram SP4 is executed according to the flowchart shown in FIG. 10. First, in step S51, it is determined whether or not the present calling of the subprogram SP4 is from the main routine.

As described above, since the present call is from step S9 of the main routine, step S52 is executed next, and a blocking process for collecting a plurality of created output data is performed. This is repeated every time a record is read from the input file FI1, and when a block consisting of a predetermined number of data is completed, steps S53 to S54 are executed, and the block is written to the output file FO1 via the external device. Then, in step S55, the number of written blocks is incremented by one.

The number of blocks to be written is the number of blocks to be written to the output file FO1 recorded in the checkpoint table TB. If the execution of this job step is "restart", the checkpoint table is written. It will be incremented by 1 with respect to the number of write blocks read from TB.

Next, in the main routine, step S10
Is executed and the corresponding record in the business database DB1 is updated based on the input record.

As described above, steps S6 to S10 of the main routine are repeatedly executed for each record of the input file FI1 while intermittently writing to the output file FO1, and at predetermined intervals during that time. Perform database synchronization processing.

That is, in step S6, as described above, the input file reading subprogram SP3 is called to execute the subroutine of FIG. 9, and in step S43 of this subroutine, the synchronization processing subprogram S is executed.
It calls P6 and executes the synchronous processing subroutine.

This synchronization processing subroutine is performed according to the flowchart shown in FIG.
Then, the number of times the synchronous processing subprogram SP6 is called is counted, and it is determined in step S62 whether or not the predetermined number of intervals has been reached. If the number of intervals has not yet been reached, then step S63 is performed.
Then, it is determined whether or not the time measured by using the timer 16 shown in FIG. 1 is equal to or longer than a predetermined interval time.

Then, when the number of times the subprogram SP6 is called reaches a predetermined number of intervals, or when the elapsed time reaches a predetermined interval time, first in step S64, the output file writing subprogram SP4 is called, The writing process to the output file FO1 is performed by the above-mentioned output file writing subroutine.

In this case, since the subprogram SP4 is called from the synchronous processing subroutine, the output file writing subroutine executes steps S51 to S56 in FIG. Determine whether there is data. Then, if there is data that has not been written by executing step S9 of the main routine, that data is set as one block and is written to the output file FO1 in step S54.
Count up the number of write blocks by 1.

Next, in the synchronization processing subroutine, as shown in FIG.
In step S65, the various data in the checkpoint table TB, that is, the number of inputs for the input file FI1, the number of blocks written to the output file FO1, the totalized result, and the like are updated based on the values in the memory 15.

Then, in step S66, the synchronous processing for the business database DB1 and the checkpoint table TB updated as described above is simultaneously executed,
Then, in step S67, the number of calls and the elapsed time of the synchronization processing subprogram SP6 are cleared to prepare for counting the number of calls and the elapsed time for the next synchronization processing.

As described above, it is assumed that the records of the input file FI1 are sequentially read in step S44 of the input file reading subroutine while performing the synchronization processing at the predetermined intervals, and the last record is read. When step S44 is executed, there is no record to be read, so that it is determined in step S7 of the main routine that the reading of all records of the input file FI1 has been completed.

If the end of reading is determined, the file close subprogram SP2 is called in step S11 of the main routine, the file close subroutine shown in the flowchart of FIG. 12 is executed, and the input / output used in the job step is executed. Close the file.

In this file close subroutine,
In step S71, the type of file to be closed is determined. If the file is an input file, the input file is closed in step S72.

If it is an output file, first, in step S73, the above-mentioned output file writing subprogram SP4 is called. In this case, this subprogram SP
Since the call of No. 4 is from the file close subroutine, as in the case of being called in the synchronous processing subroutine, step S56 of FIG. 10 is executed. If there is unwritten data, the unwritten data is written in steps S54 and S55. The data is written as one block in the output file FO1 via the external device, and the number of write blocks is incremented by one.

In other words, the data left unwritten by either step S9 of the main routine or step S64 of the synchronization processing subroutine is written to the output file FO1 at the time of file closing. Then, step S74 is executed to close the output file. Then, in step S75, it is determined whether or not there is a file to be closed next, and steps S71 to S74 are executed until the closing process is completed for all files.
Is repeatedly executed, and when all files are closed, the process returns to the main routine.

In the main routine, next, step S12
Then, the recovery / restart subprogram SP5 is called, and the subroutine of FIG. 7 is executed again. In that case,
This time is the processing at the end of the job step. Therefore, in the recovery / restart subroutine, step S
21. Steps S28 and S29 are executed. First, in step S28, "normal end" is written in the status record column of the checkpoint table TB, and step S28 is executed.
At 29, the synchronization processing is performed on the portions of the checkpoint table TB and the business database DB1 that have been updated after the previous synchronization processing was executed.

As described above, one job step of the batch job ends, but during execution of this job step, the system may end abruptly due to an abnormality or the like. The operation at the subsequent restart will be described.

Now, for example, the job step ST3 of JOB-A shown in FIG. 5 (B) reads the 120th record of the input file FI1, for example, and ends halfway when the update of the business database DB1 based on this is finished. I shall. Further, before this midway end, it is assumed that the business database DB1 has been synchronously processed when the number of input records from the input file FI1 is “100”, and the number of blocks written to the output file FO1 at that time is “ 50 ", and the total result is" 10 ". And it shall restart in this state.

At this time, the business database DB1 is rolled back to the state when the synchronous processing is performed before the midway termination, that is, the number of input cases from the input file FI1 is 100. In steps S64 and S65 of the synchronization processing subroutine shown, the checkpoint table TB is updated in the updated state at the time of synchronization processing of the business database DB1, and the checkpoint table TB and the business database DB are updated.
The synchronous processing of 1 is performed simultaneously.

Therefore, the checkpoint table T
B will also be rolled back until the number of input cases from the input file FI1 is 100, and as shown in FIG. 5B, the record of the job step ST3 of the checkpoint table TB is from the input file FI1. In the input number column of “100”, the write block number column in the output file FO1 is “50”, and the total result column is “1”.
It is recorded as "0".

As a result, when the job step ST3 is restarted, S2 of the recovery / restart subroutine of FIG. 7 is executed.
When reading the total result of the user-defined area in 6,
"10" is read as the initial value of the aggregation process, and in the positioning process of the input file FI1 in step S34 of the file open subroutine of FIG.
The restart position is set to “101”, and the restart position is set to “5” in the positioning process of the output file FO1 in step S38.
1 ”.

As a result, at the time of restart after halfway, the input file reading process of step S6 of the main routine, the totaling process of step S8, the writing process to the output file of step S9, and the operation database DB1 of step S10 The update processing and the like are all performed from the continuation of the state in which the synchronous processing is performed before the end of the processing.

In this way, the business database DB1 and
The checkpoint table TB configured as a database that records the processing history is synchronously processed in the updated state at the same time point, which may cause data duplication processing or omission when restarting after halfway. It will be restarted without any operation, and the continuity before and after the midway end of the business database DB1 will be secured.

If the job step ends normally from a normal initial start, or if it ends normally by restarting after a halfway end, a checkpoint is entered in step S28 of the recovery / restart subroutine in FIG. Since “normal end” is recorded in the status recording column of the table TB and this is synchronously processed in step S29, when a plurality of job steps constituting the batch job are restarted from the first one after the middle end, As described above, the normally ended job step is immediately ended through steps S2 to S12 of the main routine of FIG. This avoids an operation such as reading an unnecessary input file for a job step that has ended normally, and restarts a job step that has ended abruptly.

Next, as a second embodiment, the database is collectively updated at the time of a key break of the input file,
A case will be described in which an application program of a type that performs synchronous processing every predetermined number of times is used. For the sake of convenience, the database and the input / output file will be described using the same ones as in the first embodiment.

The flow chart of the main routine of the job step executed by this application program is shown in FIG. 13. The subprogram called by this main routine is the subprogram called by the main routine of the first embodiment described above. SP1-S
It is the same as P6, and these subprograms SP1
.About.SP6 are commonly used by both application programs.

Also in this second embodiment, the job step calls the recovery / restart subprogram SP5 in step S101 of the main routine of FIG. 13 after initializing all variables and the like used in the following control. Starting from that, step S2 of the subroutine of FIG.
In step 1, it is determined whether the current processing operation is the processing at the start of the job step or the processing at the end of the job step. This determination is made depending on whether or not the subprogram SP5 is called in step S101 of the main routine,
When called in step S101 of the main routine, it is determined that the current process is a start process, and otherwise, it is determined that the process is an end process.

Since the process this time is the start process,
The checkpoint table TB is referred to in step S22 of the subroutine 7 and, in step S23, depending on whether or not a record of the job step is created in this table TB, the start processing of this time is the initial start or after the halfway end. Is restarted.

Then, in case of initial start, step S2
In step 4, the name of the job step that started the process this time is written in the job step name column of the checkpoint table TB, and the status of this job step is set to "initial start" in step S25, and this is set in the memory 15
To record. In this case, checkpoint table TB
The status record column of the relevant job step is held blank.

When a record for the job step is already created in the checkpoint table TB, as in job step ST4 of JOB-A shown in FIG. 14, the current start is restarted after the midway end. Then, in step S26, the data such as the counting result recorded in the user-defined area of the record is read out onto the memory 15.

In this case, in this job step, a record record field at the time of key break is provided in the user-defined area of the checkpoint table TB as shown by the symbol c in FIG. Since the content of the record next to the last record subjected to the synchronous processing before the end is recorded, the content of this record is also read into the memory 15.

Then, in step S27, the status is set to "restart" and this is recorded in the memory 15. Even in this case, the status record column of the job step is held blank.

When the job step start processing is completed as described above, the operation returns to the main routine, and FIG.
In step S102 of the flowchart, the status of the job step is read from the checkpoint table TB, and it is determined whether the status is “normal end”. If it is not "normal end", that is, if it is "initial start" or "restart", the file open subprogram SP1 is called in step S103, and the file open subroutine shown in FIG. 8 is executed.

In this subroutine, first, step S
At 31, it is determined whether the file to be opened first according to a predetermined order is the input file,
If it is an input file, it is opened in the input mode in step S32, and then it is determined in step S33 whether the status of the job step recorded in the memory 15 is "initial start" or "restart". In the case of normal "initial start", it is determined in step S39 whether or not there is a file to be opened next.

If the status recorded in the memory 15 is "restart", the restart position is positioned in step S34. That is, when the opened file is, for example, the input file FI1 described above, the number of inputs from the input file FI1 of the job step to be processed this time recorded in the checkpoint table TB is read, and the value next to that value is input. It is positioned at the restart position of the file FI1 and recorded in the memory 15.

On the other hand, when the file to be opened this time is an output file, steps S31 to S35 are executed, and the status of the job step recorded in the memory 15 is "initial start" or "restart". To determine. And in the case of "initial start",
In step S36, the output file is opened in the output mode for newly creating, and in the case of "restart", step S3
Open the existing output file in 7 in I / O mode.

In the case of "restart", after the input / output mode is opened, the restart position is positioned in step S38. That is, when the opened file is, for example, the above-mentioned output file FO1, the number of blocks written to the output file FO1 of the job step to be processed this time recorded in the checkpoint table TB is read, and the value next to that value is read. Output file FO1
At the restart position and is recorded in the memory 15.

In any case, after opening the first file, it is determined in step S39 whether there is a file to be opened next. If there is the next file, the steps S31 to S38 are executed. Is repeatedly executed to open all the files including the input file FI1 and the output file FO1 necessary for the processing of this job step.

When the file open subroutine is completed as described above, it is then determined in step S104 of the main routine in FIG. 13 whether the current start processing is "initial start" or "restart", and "initial start" is performed. In this case, in step S105, the work area used for recording the status, recording the number of input / output cases, recording the totalized result, etc., provided on the memory 15 is initialized, and in step S106, the input file reading subprogram SP3 To execute an input file reading subroutine whose flowchart is shown in FIG.

In this subroutine, first, in step S4
In step 1, it is determined whether or not the mode of the synchronization processing for the database of the job step is the automatic synchronization mode based on the attribute of the application program. The current job step is the key every predetermined number of times of the input file FI1. Since this is a mode in which synchronous processing is performed at the time of a break, steps S42 and S43 are not executed, and records are read from the input file FI1 and the input count is counted up in steps S44 and S45. In this case, since it is the "initial start", the first record of the input file FI1 is read, and the number of input cases is "1".

On the other hand, in the case of "restart", steps S105 and S106 of the main routine are not executed and the read processing from the input file is passed. In this case, as described above, the recovery / recovery of FIG. In the restart subroutine, checkpoint table TB to memory 1
5, the aggregation result and the contents of the record next to the last record that was synchronized last time are read out.
In the file open subroutine, the positioning process of the input file FI1 and the output file FO1 is performed.

Next, step S107 of the main routine.
Then, it is determined whether or not the reading of all the records of the input file FI1 is completed. This determination is made based on whether or not a record is actually read when step S44 of the input file reading subroutine is executed in the case of "initial start". However, the record is always read immediately after "initial start". If it is determined that the reading has not ended, and if it is “restart”, step S10
Even when the reading process from the input file 6 is not performed, it is determined that the reading is not completed.

Therefore, in any case, immediately after the start, steps S108 and thereafter are executed, and first, step S1.
At 08, the aggregation process 1 is performed. This tabulation process 1 is a process of tabulating all read records by category when the input file FI1 is configured as shown in FIG. 4, for example. This is a process of executing an operation each time a record (including a record read from the user-defined area of the point table TB) is read.

Then, in step S109, the data editing area of the memory 15 is updated based on the record read from the input file FI1 and the record read from the user-defined area of the checkpoint table TB. In this data editing area, data having the same fields as the customer master shown in FIG. 3 is created for the corresponding key (customer), and the input file FI is input.
1 or one of the fields is updated by the record read from the user-defined area of the checkpoint table TB.

Then, in step S110, after performing other processing required in the job step, in step S111, the output file writing subprogram SP4 is called, and in accordance with steps S51 to S55 of the subroutine of FIG. Output data blocking, output file FO at the time when blocking is completed
Processing such as writing a block to 1 and counting up the number of writing blocks is performed.

Then, in step S112, it is determined whether or not a key break has occurred according to the record read from the input file FI1. That is, when the input file FI1 is configured as shown in FIG. 4, when the customer number, which is a key item, changes, that is, in the case of the example shown in FIG.
It is determined that a key break has occurred when the fifth or 103rd record is read.

Immediately after "initial start" or "restart" in which only one record is read, it is not judged that a key break has occurred. In this case, step S117 is executed and the input file reading subprogram SP3 is executed. Is called again and the subroutine of FIG. 9 is executed again. And
Until it is determined in step S107 that all the records have been read, the records are read by the input file reading subroutine, the number of input cases is counted, the totaling process 1 by the main routine, the data editing area is updated, and the writing process to the output file is performed. Etc. are repeatedly executed.

If a key break is determined in step S112 while the above processing is repeatedly executed for the sequentially read records, the next step S
113 to S116 are executed, and first, the aggregation process 2 is executed in step S113. This tabulation process 2 shows the number of customers who have updated the data in any of the categories, in other words, the number of keys that are the update targets, when the input file FI1 is configured as shown in FIG. It is a process of counting up, and is a process of counting up for each key break.

Next, in step S114, the data before the key break created in the data editing area of the memory 15,
That is, the corresponding record of the business database DB1 is updated based on the updated data of each item of the customer before the key break. Then, in step S115, the synchronization processing subprogram SP6 is called, and the synchronization processing subroutine shown in FIG. 11 is executed.

This synchronization processing subroutine is executed for each key break, but the actual synchronization processing is step S6.
When it is determined in 2 that the number of times the subprogram SP6 is called is equal to or greater than the predetermined interval number, or when it is determined in step S63 that the elapsed time is equal to or greater than the predetermined interval time. To be done.

Then, when the number of times the subprogram SP6 is called is equal to or greater than the interval number, or when the elapsed time is equal to or greater than the interval time, the output file writing subprogram SP4 is first called in step S64. , Output file FO
Write processing to 1.

This processing is executed according to the subroutine shown in the flow chart of FIG. 10. Since the calling of the output file writing subprogram SP4 this time is from the synchronous processing subroutine, steps S51 to S56 are executed in the subroutine of FIG. Then, it is determined whether or not there is unwritten data in the output file FO1. If there is data that has not been written in step S111 of the main routine,
With those data as one block, step S54,
According to S55, writing to the output file FO1 and counting up of the number of writing blocks are performed, and the process returns to the synchronization processing subroutine.

In the synchronous processing subroutine, after the above-mentioned writing processing is performed, in step S65, the checkpoint table is calculated using the number of records input at that point, the number of writing blocks, or the counting results of the counting processing 1 and 2. Update TB. In this case, in this job step, the contents of the record at the time of the key break that calls the synchronous processing subprogram SP6 this time are recorded in the record record column at the key break provided in the user-defined area of the checkpoint table TB. .

Then, in step S66, the synchronous processing is simultaneously executed for the checkpoint table TB and the business database DB1 updated as described above,
Then, in step S67, the counted number of times the subprogram SP6 is called and the measured elapsed time are cleared.

Here, for example, as shown in FIG. 4, it is assumed that a key break occurs at the 103rd record of the input file, and the number of intervals or the interval time for executing synchronous processing has been reached at that time. Then, the business database DB1 is 1 immediately before the key break.
The synchronization processing is performed in the updated state with the 02nd record, and as shown in FIG. 14, the contents of the 103rd record are stored in the record record column at key break in the user-defined area of the checkpoint table TB. It is recorded, and "103" is recorded in the input number record column. Further, as the number of write blocks and the results of the aggregation processes 1 and 2, the values when the 103rd record is processed are recorded. Then, in this state, the checkpoint table TB is also synchronized.

Then, returning to the main routine, the data recorded in the data editing area of the memory 15 is erased in step S116 to prepare for the data of the next key.

Next, in the input file reading process of step S117, the second record from the key break, that is, the sixth record and the 104th record in the example of FIG. 4, is read, and this record and step S117 are performed thereafter.
By the update processing of the data editing area in step S109 based on the record read in step S1 and the record immediately after the key break read out from the user-defined area of the checkpoint table TB, that is, the fifth and 103rd records in FIG. , Create data for the following keys. Then, at the next key break, update processing or synchronization processing of the business database DB1 based on the data is executed in the same manner as in the above case.

As described above, it is assumed that the records of the input file FI1 are sequentially read in step S44 of the input file read subroutine and the last record is read while performing the synchronization processing at the key breaks every predetermined number of times. In the main routine, it is determined in step S107 that reading of all the records of the input file FI1 has been completed. And then step S
At 118, the file close subprogram SP2 is called, and the file close subroutine shown in the flowchart of FIG. 12 is executed.

In this subroutine, first, in step S7
The type of file to be closed is determined in 1 and if the file is an input file, the input file is closed in step S72. If it is an output file, first, step S73.
Then, the output file writing subprogram SP4 described above is called. In this case, since the subprogram SP4 is called from the file close subroutine,
Similar to when called in the synchronous processing subroutine, step S56 of FIG. 10 is executed, and if there is unwritten data, the unwritten data is output as one block via an external device in steps S54 and S55. While writing to the file FO1, the number of write blocks is counted up.

That is, step S11 of the main routine.
The data left unwritten by any one of step 1 and step S64 of the synchronization processing subroutine is written to the output file FO1 at the time of file closing. Then, step S74 of the file close subroutine is executed to close the output file.

Then, in step S75, it is determined whether or not there is a file to be closed next, and steps S71 to S74 are repeatedly executed until the closing processing is completed for all files, and when all the files are closed, the main routine is executed. Return.

Next, step S119 of the main routine.
Then, the business database DB1 is updated based on the data about the last key created in the data editing area of the memory 15 after the last key break. Then, in step S120, the recovery / restart subprogram SP5
Again to call step S2 of the subroutine in FIG.
From step 1 to step S28, "normal end" is written in the status record column of the checkpoint table TB, and at step S29, the checkpoint table TB is synchronously processed, and at the same time, the business database DB1.
As for the above, the synchronization processing is performed in the state updated in step S119.

As described above, this job step ends. Next, the operation at the time of restarting when this job step ends halfway will be described.

Now, for example, the job step ST4 of JOB-A shown in FIG. 14 is, for example, 12 of the input file FI1.
It is assumed that the 0th record is read and the operation database DB1 and the checkpoint table TB based on the 0th record are updated, and the processing is terminated halfway due to an abnormality or the like. In addition, before the end of this process, the business database D
It is assumed that B1 has been synchronously processed with the number of input records from the input file FI1 being updated based on the records up to the 102nd record at the time of the 103rd key break. And it shall restart in this state.

At this time, the business database DB1 is rolled back to the state when the synchronous processing is performed before the midway termination, that is, the state when the number of inputs from the input file FI1 is 102. At the time of executing the synchronous processing of the business database DB1, in steps S64 and S65 of the synchronous processing subroutine, the checkpoint table TB is also updated in the state at that time and the synchronous processing is performed. Therefore, as shown in FIG. In the point table TB, the number of inputs is recorded as "103", and the content of the 103rd record is recorded in the record record field at key break in the user-defined area.
Further, the values at the time when the number of input cases is 103 are also recorded in the number of writing blocks and the result of the aggregation processing 1 and 2.

Therefore, when the job step ST4 is restarted, S of the recovery / restart subroutine of FIG. 7 is executed.
When the data of the user-defined area of the checkpoint table TB is read at step 26, the contents of the 103rd record and the aggregation result at the time when the 103rd record is read will be read out. In the positioning process for the input file FI1 in step S34 of the file open subroutine, the restart position is set to the 104th record, and in the positioning process for the output file FO1 in step S38, the write block at the time when the number of inputs is 103 The next value of the number is the restart position.

If the restart after the midway termination is executed in this state, the 104th record of the input file FI1 is read in step S117 of the main routine. At this time, the recovery / restart is performed. At the step S26 of the subroutine, the 103rd record is read from the user-defined area of the checkpoint table TB.
The processing after the middle end is started from the 03rd record.

Further, the counting of the number of writing blocks and the totaling processing 1, 2 etc. are restarted with the value at the time when the number of input records is up to 103 as the initial value. The synchronization process will be continued from the key next to the key after the synchronization process without duplication or omission of the process, and the continuity of the process before and after the abrupt termination will be ensured.

Also in this embodiment, the recovery / restart subroutine of FIG. 7 is used when the job step ends normally from the normal initial start, and when the job step ends normally by restarting after the midway end. In step S28 of the above, "normal end" is recorded in the status record column of the checkpoint table TB.
Since the synchronous processing is performed in step 29, when a plurality of job steps that constitute the batch job are restarted from the first job step after the midway termination, as for the job step that normally ends as described above, the main routine of FIG. Immediately after step S102 to step S120, the process ends.

Next, as a third embodiment, an application program of a type that performs output processing such as printout and other processing based on the record recorded in the business database DB1 without using an input file. The case of use will be described.

Also in this embodiment, except for the input file reading subprogram SP3, the subprogram SP used in the above first and second embodiments is used.
1 to SP6 are used. Further, for convenience, the same database and files as those in the first embodiment will be used for description.

The main routine of the job step executed by this application program is shown in FIG.
The process is executed according to the flowchart shown in step S1, and first, all variables used in the following control are initialized, and then step S2
At 01, the recovery / restart subprogram SP5 is called, and the subroutine shown in the flowchart in FIG. 7 is executed.

In this subroutine, firstly in step S2
In step 1, it is determined whether the current processing operation is the processing at the start of the job step or the processing at the end of the job step. This determination is made depending on whether or not the subprogram SP5 is called in step S201 of the main routine, but since it is called in step S201 of the main routine this time, it is determined to be start processing. Then, in step S22, the checkpoint table TB is referred to, and in step S23, whether or not the start process of this time is the initial start depending on whether or not the record of the job step is created in the table TB. It is determined whether it is restart after the end.

In the case of initial start, in step S24, the name of the job step that started the processing this time is written in the job step name column of the checkpoint table TB, and the status of this job step is set to "initial" in step S25. “Start”, and this is recorded in the memory 15. In this case, the status record column of the job step in the checkpoint table TB is held blank.

Further, as in the job step ST5 of JOB-A shown in FIG. 16, when the record of the job step is already created in the checkpoint table TB, the start of this time is restarted after the halfway end. Then, in step S26 of the recovery / restart subroutine, the data such as the counting result recorded in the user-defined area of the record is read out onto the memory 15.

In this case, in this job step, an updated key value recording column for recording the key value of the updated record in the business database DB1 in the user-defined area of the checkpoint table TB, as indicated by the symbol d in FIG. As described later, since the key value of the last record that has been synchronously processed before the end of the processing is recorded in this column, this key value is also read out on the memory 15.

Then, in step S27, the status is set to "restart" and this is recorded in the memory 15. Even in this case, the status record column of the job step is held blank.

When the job step start processing is completed as described above, the operation returns to the main routine, and the operation shown in FIG.
In step S202 of the flowchart of FIG. 5, the status of the job step is read from the checkpoint table TB, and it is determined whether the status is “normal end”. If it is not "normal end", that is, if it is "initial start" or "restart", the file open subprogram SP1 is called in step S203, and the file open subroutine shown in FIG. 8 is executed.

In this subroutine, first, in step S
In step 31, it is determined whether or not the first file to be opened is the input file according to the predetermined order. However, since the input file is not used in this job step, only the output file is targeted. S35 is executed to determine whether the status of the job step recorded in the memory 15 is "initial start" or "restart".

Then, in the case of "initial start", the output file is opened in the output mode in which a new output file is created in step S36, and in the case of "restart", the existing output file is opened in the step S37. To open the file. Further, in the case of "restart", after opening in the input / output mode, positioning processing of the restart position is performed in step S38. That is, if the opened file is, for example, the output file FO1 described above,
The number of blocks written to the output file FO1 of the job step to be processed this time, which is recorded in the checkpoint table TB, is read, the value next to that value is positioned at the restart position of the output file FO1, and is recorded in the memory 15.

Thereafter, in step S39, it is determined whether or not there is a file to be opened next. If there is a next file, steps S31 and S35 to S38 are repeatedly executed to process the job step this time. Open all required files with.

When the file open subroutine is completed as described above, it is then determined in step S204 of the main routine of FIG. 15 whether the current start processing is "initial start" or "restart", and "initial start" is performed. In the case of "," in step S205, the work areas for recording various data provided on the memory 15 are initialized, and in step S2
In 06, the extraction condition of the business database DB1, that is,
The key value of the record extracted from the business database DB1 is initialized, and the value is set to "0".

In the case of "restart", step S2
The updated key read from the user-defined area of the checkpoint table TB to the memory 15 by the recovery / restart subroutine of FIG. Set the value in the extraction condition of the business database. Regarding the output file FO1,
Positioning processing is performed by the file open subroutine of FIG.

Next, step S208 of the main routine.
Then, one record to be processed is extracted from the business database DB1. At this time, the records to be extracted are
In the case of the record of the key value next to the key value set as the extraction condition, and in the case of “initial start”, the key value as the extraction condition is “0”, so the record with the key value of 1, That is, the first record recorded in the business database DB1 will be extracted, and in the case of "restart", the key value next to the updated key value set as the extraction condition in step S207 will be displayed. Records will be extracted.

Then, in step S209, it is determined whether or not the extraction of all the records to be processed is completed. However, since the extraction of all the records is not completed at the beginning, step S210 is executed next and the job step Performs the aggregation and other processing required in. After that, in step S211, the output file writing subprogram S
P4 is called and step SS of the subroutine of FIG.
In accordance with 1 to S55, processing such as blocking of output data, writing of a block to the output file FO1 at the time when blocking is completed, and counting up of the number of writing blocks is performed.

Next, in step S212, it is indicated that the processing for the record is completed in the business database DB.
Then, the business database DB1 is updated, and then the synchronous processing subprogram SP6 is called in step S213 to execute the synchronous processing subroutine shown in FIG.

In this synchronization processing subroutine, the number of times the subprogram SP6 is called is counted in step S61, and when it is determined in step S62 that the number is equal to or greater than the predetermined interval number, or in step S63. When it is determined that the elapsed time is equal to or longer than the predetermined interval time, the processing of step S64 and thereafter is executed.

That is, first, in step S64, the output file writing subprogram SP4 is called, and writing processing to the output file FO1 is performed according to the subroutine whose flowchart is shown in FIG.

In this case, since the calling of the output file writing subprogram SP4 this time is from the synchronous processing subroutine, in the subroutine of FIG. 10, steps S51 to S56 are executed, and at this time, the unwriting to the output file FO1 is not performed. Determine whether there is data. Then, if there is data for which writing has not been completed by executing step S211 of the main routine, those data are regarded as one block, and step S2 is performed.
The data is written to the output file FO1 at 54, the number of write blocks is counted up at step S55, and the process returns to the synchronization processing subroutine.

Then, in the synchronous processing subroutine, in step S65, the checkpoint table TB is updated using the number of blocks to be written to the output file FO1 at that time, or the totalization result, and at that time, as described above. The key value of the last record of the business database DB1 that has completed processing is recorded in the updated key value recording field provided in the user-defined area of the checkpoint table TB.

Thereafter, this checkpoint table T
B and the business database DB1 are simultaneously synchronized. Then, in step S67, the counted number of times the subprogram SP6 is called and the measured elapsed time are cleared.

As described above, while performing synchronous processing at a predetermined interval, each record to be processed in the business database DB1 is sequentially extracted and the predetermined processing is performed.
Assuming that the last record has been completed, there is no record to be read the next time step S208 is executed, so it is determined in step S209 that extraction and processing have been completed for all records.

If the end is determined, the file close subprogram SP2 is called in step S214, the file close subroutine shown in the flowchart of FIG. 12 is executed, and the output file used in the job step is closed.

In this subroutine, first, step S7
The type of file to be closed is determined in 1, but since only the output file is present this time, the next step S73
To call the output file writing subprogram SP4.

In this case, since the subprogram SP4 is called from the file close subroutine, the step S56 of FIG. 10 is executed as in the case of being called in the synchronous processing subroutine, and when there is unwritten data. In steps S54 and S55, the unwritten data is written as one block in the output file FO1 via the external device, and the number of written blocks is counted up.

That is, step S21 of the main routine.
The data left unwritten by any one of step 1 and step S64 of the synchronization processing subroutine is written to the output file FO1 at the time of file closing. Then, step S74 is executed to close the output file.

Then, in step S75, it is determined whether or not there is a file to be closed next, and steps S71 to S74 are executed until the close processing is completed for all files.
Is repeatedly executed, and when all files are closed, the process returns to the main routine.

Next, in step S215, the recovery / restart subprogram SP5 is called again, and the recovery / restart subprogram SP5 shown in FIG.
Step S21 to Step S28 of the subroutine is executed, "Normal end" is written in the status record column of the checkpoint table TB, and Step S29 is executed.
Then, the synchronization processing is performed in a state in which the checkpoint table TB and the business database DB1 are updated after the previous synchronization processing is executed.

As described above, this job step ends. Next, the operation at the time of restart when this job step ends midway due to occurrence of an abnormality or the like will be described.

Now, for example, it is assumed that the job step ST5 of JOB-A shown in FIG. 16 is terminated abruptly, and the job database DB1 has the key value "250" before this abrupt termination.
0 ", the number of writing blocks to the output file FO1 is" 2 "
It is assumed that the synchronous processing has been performed at the time of "00". Therefore, when restarting, the business database DB1 is rolled back to the record with the key value “2501”.

At this time, the checkpoint table TB
Also, since the synchronous processing is performed in the state when the business database DB1 is synchronously processed, the number of writing blocks is recorded as "200" as shown in FIG. 16, and the updated key value recording column of the user-defined area is recorded. "2500" is recorded. In addition, when aggregation processing is performed,
The result of the aggregation is also the key value “2 of the business database DB1.
The value at the time of updating the record "500" is recorded.

Therefore, when the job step ST5 is restarted, the recovery / restart subprogram SP5 is called in step S201 of the main routine, as shown in FIG.
When the data of the user-defined area is read in S26 of the subroutine of, the key value of the updated record is "2.
When "500" is read, its key value "250"
The number of blocks to be written and the totalization result when the record of "0" is updated are read.

In addition, step S203 of the main routine.
Then, when the file open subprogram SP1 is called and the subroutine of FIG. 8 is executed, step S
In the positioning processing of the output file FO1 of 38, the next value of the write block number “200” is positioned at the restart position.

As a result, step S2 of the main routine
The key value "2500" is set as the extraction condition in 07, and the next key value "250" is set in step S208.
The record of "1" will be extracted. Further, in step S210, the writing process to the output file FO1 is performed as the “201st” item, and in this way, continuity is maintained before and after the abrupt termination and based on the record of the business database DB1. Processing will be performed.

Also in this embodiment, the recovery / restart subroutine of FIG. 7 is used when the job step ends normally from the normal initial start, and when the job step ends normally due to restart after abnormal end. In step S28 of the above, "normal end" is recorded in the status record column of the checkpoint table TB.
Since the synchronous processing is performed in step 29, when a plurality of job steps constituting the batch job are restarted from the first one after the abnormal end, as for the job step normally ended as described above, the main routine of FIG. Immediately after step S202 to step S215, the process ends.

[0178]

As described above, according to the batch processing system of the present invention, during the execution of the batch job using the business database, the synchronous processing is performed on the business database at a predetermined timing, and at the same time, Since the progress management table configured as a database that manages the progress status of jobs is updated and the synchronization processing is performed, when restarting after a halfway termination due to an error etc. By being rolled back to the state, the recorded contents and the recorded contents of the progress management table become the same point in time.

Therefore, if the job is restarted based on the progress management information recorded in the progress management table after the end of the process, duplication or omission of processing will not occur and the recorded contents and output files of the business database will not occur. Continuity will be maintained before and after the end of the data.

Further, according to the invention of claim 15 of the present application, the program for operating the system is a main program for executing the original batch job and a recovery / recovery program.
Since it is divided into subprograms for the restart function, it is possible to easily provide a recovery / restart function to an existing application program for batch jobs. As a result, the development cost can be reduced, and the sub program can be shared by different main programs, so that the entire program can be simplified.

[Brief description of drawings]

FIG. 1 is a diagram showing a hardware configuration of a system according to an embodiment of the present invention.

FIG. 2 is a diagram showing a configuration of a program of the system.

FIG. 3 is a diagram showing a configuration of a business database in the system.

FIG. 4 is a diagram showing a structure of an input file in the same manner.

FIG. 5 is a diagram showing the structure of a checkpoint table.

FIG. 6 is a flowchart showing a main routine according to the first embodiment of the present invention.

FIG. 7 is a flowchart showing a recovery restart subroutine.

FIG. 8 is a flowchart showing a file open subroutine.

FIG. 9 is a flowchart showing an input file reading subroutine.

FIG. 10 is a flowchart showing an output file writing subroutine.

FIG. 11 is a flowchart showing a synchronization processing subroutine.

FIG. 12 is a flowchart showing a file close subroutine.

FIG. 13 is a flowchart showing a main routine according to a second embodiment of the present invention.

FIG. 14 is an explanatory diagram of a checkpoint table during implementation of this embodiment.

FIG. 15 is a flowchart showing a main routine according to a third embodiment of the present invention.

FIG. 16 is an explanatory diagram of a checkpoint table during implementation of this embodiment.

[Explanation of symbols]

DB1 Business database TB Progress management table (checkpoint table) FI1 input file FO2 output file AP main program (application program) SP1 to SP6 subprograms

Continued front page    F-term (reference) 5B027 AA01 AA04 BB01 CC04                 5B082 DC08 DC09                 5B085 AC17 BA01                 5B098 AA03 GA03 GB01 GC01 GD03                       GD14

Claims (16)

[Claims] 1. A batch processing system for executing a batch job using a business database by a batch job execution means, wherein a progress management table configured as a database for recording progress management information of the batch job, and execution of the batch job. At a predetermined time, the progress management information is recorded in the progress management table, and the synchronization processing means performs the synchronization processing for the progress management table and the synchronization processing for the business database at the same time; Based on the recorded contents of the progress management table, a start state determination means for determining whether the current start is an initial start or a restart after a halfway end, and when the determination means determines restart, Based on the recorded progress management information, the state when the synchronization process was performed before the end A batch processing system having a recovery / restart function, comprising: a start position locating means for locating a start position of data to be processed so that a batch job is restarted as an initial state. 2. In a system in which a batch job execution means executes a batch job based on an input file to a business database, the synchronization processing means has a predetermined interval based on the start time of the job or the previous execution time of the synchronization processing. Then, the number of records input from the input file is recorded as progress management information in the progress management table, and the progress management table and the business database are synchronized in the updated state at that time, and the start position locating means is used for the progress management. The next record of the number of inputs recorded in the table is positioned at the start position of the input file, and the batch job execution means starts the job from the record positioned at the start position by the above positioning means in the input file when the job is restarted. The method according to claim 1, wherein the execution is performed. Batch processing system with recovery and restart functions. 3. In a system in which a batch job execution means executes a batch job based on an input file to a business database, the synchronization processing means stores a progress management table when a key break is detected every predetermined number of times in the input file. As the progress management information, the number of records input from the input file and the contents of the record when the key break is detected in the input file are recorded, and the synchronous processing of the business database is performed in the updated state by the record immediately before the key break. Synchronize the management table in the updated state at the time of detecting the key break,
The start position locating means positions the next record of the number of inputs recorded in the progress management table at the start position of the input file, and the batch job execution means, when the job is restarted, the key break from the progress management table. The recovery / restart function according to claim 1, wherein the record at the time of detection is read, and the record of the input file is read from the record positioned at the start position by the positioning means to execute the job. Batch processing system. 4. In a system in which a batch job execution means executes a batch job based on a record in a business database, the synchronization processing means records a key value of a processed record in the business database as progress management information in a progress management table. The start position locating means positions the record of the key value next to the key value of the processed record recorded in the progress management table at the start position of the business database, and the batch job executing means determines when the job is restarted. 2. The batch processing system with a recovery / restart function according to claim 1, wherein the job is executed from the record positioned at the start position by the positioning means in the business database. 5. A system in which processing is performed on an output file by a batch job execution means, the synchronization processing means records the number of cases of writing data to the output file as progress management information in a progress management table, and a start position locating means. Positions the data next to the number of writings recorded in the progress management table at the start position of the output file, and the batch job execution means is positioned at the start position by the above-mentioned positioning means in the output file when the job is restarted. A batch processing system having a recovery / restart function according to any one of claims 1 to 4, wherein a job is executed from data. 6. The auxiliary job executing means for performing a predetermined auxiliary job during execution of a job by the batch job executing means is provided, and the progress management table includes information on the progress of the auxiliary job by the auxiliary job executing means. Is provided, and auxiliary work information reading means for reading the information in the auxiliary work information recording field as the initial value of the auxiliary work after the restart is provided when the job is restarted. A batch processing system having a recovery / restart function according to any one of claims 1 to 5. 7. A system configured to execute a batch job including a plurality of job steps for each job step, wherein a progress management table includes:
A status recording column indicating the status of each job step is provided, and a status recording unit for normalizing the status recording column of the job step when the job step ends normally is provided, and the batch job is executed from the first job step. When restarting, a job step skip means for reading the status in the status recording column of the job step at the start of each job step and skipping the job step at the normal end is provided. Item 1 to claim 6
A batch processing system equipped with the recovery / restart function described in any of 1. 8. A program for a batch processing system comprising a business database used in batch jobs and a progress management table configured as a database for recording progress management information of batch jobs, the computer comprising: Batch job executing means for executing a batch job, progress management information is recorded in the progress management table at a predetermined time during execution of the batch job by the executing means, and synchronization processing for the progress management table and the business database are performed. Synchronous processing means for performing synchronous processing at the same time, at the start of the batch job, based on the recorded contents of the progress management table, a start state determination means for determining whether the current start is an initial start or a restart after a halfway end, and When the determination means determines that the restart is to be performed, the progress management table is Based on the progress management information recorded in the table, as a start position positioning unit that positions the restart position of the processing target data so that the batch job is restarted with the state when the synchronous processing was performed before the midway end as the initial state. A program for a batch processing system equipped with a recovery / restart function that is characterized by functioning. 9. A program used in a system in which a batch job execution means executes a batch job based on an input file with respect to a business database, and when a computer functions as a synchronization processing means, at the time of starting a job or the last time. The number of records input from the input file is recorded in the progress management table as progress management information at a predetermined interval based on the execution time of the synchronization process, and the progress management table and the business database are synchronized in the updated state at that time. When performing the function to perform processing and to function as the start position locating means, the function is set to position the record next to the number of inputs recorded in the progress management table at the start position of the input file, and execute the batch job. When you want to function as a means, 9. A program for a batch processing system having a recovery / restart function according to claim 8, characterized by causing a job to be executed from a record positioned at a start position by the positioning means in an input file. 10. A program used in a system for executing a batch job based on an input file to a business database by a batch job executing means,
When making a computer function as a synchronization processing means,
When a key break is detected every predetermined number of times in the input file, the number of records input from the input file and the contents of the record when the key break is detected in the input file are recorded in the progress management table as progress management information. When the synchronization processing of the database is performed in the updated state of the record immediately before the key break and the synchronization processing of the progress management table is performed in the updated state at the time of the detection of the key break, and when it functions as the start position locating means, When the next record of the number of inputs recorded in the management table is to be positioned at the start position of the input file and also to function as a batch job execution means, when the job is restarted, the key break from the progress management table is used. Read the record at the time of detection and 9. The batch processing system having a recovery / restart function according to claim 8, wherein the batch processing system has a function of reading a record of an input file from a record positioned at the start position by the mounting means and executing a job. program. 11. A program used in a system in which a batch job execution means executes a batch job based on records in a business database, and when a computer functions as a synchronization processing means, the progress management information is stored in a progress management table as progress management information. When functioning to record the key value of the processed record of the business database and functioning as the start position locating means, the record of the key value next to the key value of the processed record recorded in the above progress management table is set. When functioning so as to be positioned at the start position of the business database and as the batch job execution means, it functions so as to execute the job from the record positioned at the start position by the above positioning means in the business database when the job is restarted. Characterized by The recovery according to claim 8
Program for batch processing system with restart function. 12. A program used in a system in which processing is performed on an output file by a batch job executing means, and when a computer functions as a synchronous processing means, a progress management table stores the progress management information to the output file. When it is made to function to record the number of data to be written out and to function as the start position positioning means, it is made to function to position the data next to the number of data to be written out recorded in the progress management table at the start position of the output file, and The function as a batch job execution means is such that when the job is restarted, the job is executed from the data positioned at the start position by the positioning means in the output file. Recovery lister described in any of A program for batch processing systems that has a function for batch processing. 13. A program used in a system in which a progress management table is provided with an auxiliary work information recording field for recording information on the progress of auxiliary work, wherein the computer executes a predetermined job during execution of the job by the batch job executing means. Auxiliary work execution means for performing auxiliary work and recording information on the progress of the auxiliary work in the auxiliary work information recording column of the progress management table, and after restarting the information in the auxiliary work information recording column when the job is restarted 13. A program for a batch processing system having a recovery / restart function according to claim 8, which functions as auxiliary work information reading means for reading as an initial value of the auxiliary work. 14. A system in which a batch job including a plurality of job steps is configured to be executed for each job step, and a progress recording table includes a status recording column indicating the status of each job step. A program used, wherein the computer restarts a batch job from the first job step, and status recording means for normally ending the status recording column of the job step in the progress management table when the job step ends normally. 9. At the time of starting each job step, the status in the status recording column of the job step is read, and when the job step ends normally, it functions as a job step skip means for skipping the job step. To any of 13 A program for a batch processing system equipped with the described recovery / restart function. 15. A main program for causing a computer to function as a batch job executing means, and a subprogram called by the main program for causing the computer to function as each means other than the batch job executing means. 15. A batch processing system program provided with the recovery / restart function according to claim 8. 16. A computer-readable recording medium in which the program according to any one of claims 8 to 15 is recorded.