JP2011258248A - System for automatically creating job network and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system for automatically creating a job network to convert a JCL file described in a JCL language for use in a mainframe into a job network flow of a plain form.SOLUTION: A system for automatically creating a job network comprises: a JCL (Job Control Language) file input unit; a job network generation unit; and a job network output unit. The JCL file input unit inputs a JCL file. The job network generation unit automatically generates a flow link based on a JCL description of the JCL file. The job network output unit converts the flow link into a job network flow and outputs the job network flow.

Description

  The present invention relates to a job network automatic generation method and program.

  In recent years, in many enterprise systems, system renewal from a mainframe to an open system using a general-purpose OS such as a UNIX (registered trademark) server or Windows (registered trademark) has been actively performed. However, in such system renewal, migration of job assets that have been operated on the mainframe until now becomes a big problem. Generally, these job assets are described in a job control language (JCL), but general-purpose job management software on an open system cannot interpret this language. Therefore, job control information expressed by these JCL descriptions is transferred to a job network flow (flow diagram) in general-purpose job management software.

  Here, there are various job network flow formats depending on each job management software. FIGS. 1-1 to 1-4 are all job network flows representing the same order relationship. In general job management software, there are many methods that allow a user to make a free description without any restriction on the description. However, in the case of a method in which no constraint is provided, even if job network flows that express the same order relationship are created, depending on the sensitivity of the operator, the completed job network flow becomes FIG. It is not certain if it will be 3. Mixing different types of job network flows with the same order relationship is problematic both in terms of test efficiency during job migration and maintainability after migration.

  In order to solve such a problem, as shown in FIG. 1-1, the starting point and the ending point are fixed in advance, and the job management software side restricts how the lines of the job network flow are drawn, thereby giving an order relationship. There is job management software that takes a method that can create an almost unique job network flow if possible. This job management software appeals that no matter who creates a job network flow, it becomes a flow in the same format and can create a simple job network flow with high maintainability.

  However, in order to migrate a JCL file onto such job management software, it is necessary to convert the JCL description into a simple job network flow in that format. Several problems can be pointed out in this conversion.

  The first problem is that when the conversion work is performed manually by a system engineer, in a user environment where a large number of JCL files exist, quality degradation due to human error and test work performed to recover it. Will consume enormous costs. Therefore, it is necessary to automate the migration work.

  The second problem is as follows. The JCL file is a special control format in which control instructions such as parallel execution and waiting are described in the actual processing description of the command interpreter-like grammar. The control command includes a control command that can be described in an arbitrary location such as \ SUBJOB⇔ ¥ ENDSUBSUBJOB (parallel execution), and a control command that clearly waits at a location described as \ WAITSUB. And exist. When such a JCL file is converted by a simple automatic conversion tool, a complicated job network flow in which crossing of lines as shown in FIG. 1-4 occurs is often generated. In this case, the created job network flow cannot be applied to the above-described job management software.

  The following inventions are known as related technologies. Japanese Patent Laid-Open No. 2001-166928 (see Patent Document 1) describes an invention of an automatic generation apparatus that automatically generates a job net. The automatic generation device creates a database for centrally managing design information from a screen for entering job net design information (system flow, job flow, same job, data item information definition, record information definition, file information definition, operation schedule). create. A job net configuration file is generated from the registered operation schedule through the job net configuration file generation mechanism. This automatic generation apparatus inputs a job network flow and does not input a JCL file.

  Japanese Patent Laid-Open No. 2001-282549 (see Patent Document 2) describes an invention of a program conversion apparatus. The program conversion device includes a plurality of program counters and a plurality of thread execution devices. The plurality of thread execution devices can execute the thread in the control speculative mode and execute the thread in the data dependent speculative mode. The control speculative mode is a mode in which instructions of a plurality of threads are simultaneously fetched, interpreted, and executed according to a plurality of program counters, and changes made to the register set after the thread generation can be canceled later. In data-dependent speculative mode, if the parent thread that created the thread stores the value in the same memory area after the thread has loaded the value from the memory area, at least the processing result of the thread after the load is discarded In this mode, these processes are re-executed.

  The program conversion apparatus converts a given source program for a multi-thread processor. The multi-thread processor has an instruction set that can execute the following as a single or a combination of at most several machine language instructions. First, the thread being executed in the thread execution device generates a new thread in the control speculative mode. Second, if the specified condition is satisfied, the own thread is terminated and the control speculative mode of the control speculative mode thread generated by the own thread is released. Thirdly, the generated thread in the control speculative mode is discarded. Fourthly, when a thread generated by the own thread loads from the memory area of the designated address, it is instructed in advance to pause the operation. Fifth, canceling the load suspension instruction for the designated memory address. Sixth, the thread executing in the thread execution device generates a new thread in the data-dependent speculative mode. Seventh, releasing the data-dependent speculative mode of the thread in the data-dependent speculative mode generated by its own thread.

  The program conversion apparatus includes a register allocation trial unit, a fork (FORK) location determination unit, an instruction rearrangement unit, and a register allocation unit. The register allocation trial unit attempts register allocation prior to parallelization, and predicts each variable in the intermediate program and the register allocation status of the intermediate term. The fork location determination unit determines whether or not to convert the conditional branch portion in the intermediate program into parallel code using a thread generation instruction based on the register allocation trial result in the register allocation trial unit, and in the parallel code Determine the parallel execution method. The instruction rearrangement unit converts the conditional branch part in the intermediate program into parallel code using a thread generation instruction based on the determination result in the fork location determination unit, refers to the register allocation trial result, An instruction for guaranteeing data dependency via the memory is inserted before and after the thread generation instruction, and the instructions before and after the thread generation instruction are rearranged so that the thread generation is performed at an early stage. The register allocation unit performs deterministic register allocation so as to obtain the same allocation result as that at the time of register allocation regarding whether or not a physical register is allocated to the parallelized and rearranged instruction sequence.

  This program conversion apparatus is one of compiler technologies, and it is considered that the main point is to avoid data dependency between processes. What is optimized is the efficiency of parallel processing.

  Japanese Patent Application Laid-Open No. 2007-257384 (see Patent Document 3) describes an invention of a job net diagram automatic generation method for automatically generating a batch system design drawing. The job net diagram automatic generation method includes a database unit and an information extraction / display unit that retrieves and displays information from the database unit. The database unit associates the job information, the program information and the file information indicating the IDs of the jobs, programs and files constituting the batch system, the job classification information and the file classification information indicating the classification of the job and the file, and the above information. Association information to be shown. The information extraction / display unit includes a program including an extracting step and a generating step. In the extracting step, job classification information is specified, and job information having the classification is extracted from the database unit. In the generating step, a graphic indicating the relationship of the job information is generated by determining the context of the job information extracted from the association information. In this job net diagram automatic generation method, it is not clear with what policy the job network flow is created based on the order relation information. Therefore, when the number of jobs is huge and the order relation is complicated, the created flow is complicated and difficult to understand, and it is highly likely that the job network flow is not practical. In addition, the possibility of being converted into a different job network flow at each conversion (as the “picture” that the flow has the same logical meaning but shown to humans) cannot be denied.

  Japanese Patent Laid-Open No. 8-147156 (see Patent Document 4) describes an invention of a program porting system. The program porting system transplants a series of programs that are automatically executed using a job control language in a porting computer system so that the program can be automatically executed on the porting computer system. The program porting system includes a job control program input process, an input file existence search process, a logical relationship review process, an external reference file input process, a log list output process, and a change program output process.

  In the job control program input process, a job control program described in a job control language for the porting computer system is input. In the input file existence search process, it is searched whether a file corresponding to the input file specified in the file specification statement of the job control program exists on the storage device in the computer system of the porting destination. In the logical relation review process, the logical relation described in the job control program is examined. In the external reference file input process, a file corresponding to the external reference file described in the job control program is input from the storage device in the ported computer system. The log list output process outputs a log list indicating the presence of an error when an error occurs in the program input process, input file existence search process, logical relationship inspection process, and external reference file input process, but no error occurred Output a log list indicating the absence of errors. In the changed program output process, if no error occurs in the program input process, input file existence search process, logical relationship inspection process, and external reference file input process, the job control program is changed and output for the destination computer system. To do.

  In this program porting system, it is not expected to optimize the simplification of the flow shape and to improve the performance of job network flow readability and visualization. Furthermore, it can be considered that it does not lead to a reduction in management man-hours, operation man-hours and maintenance man-hours of the job execution environment.

JP 2001-166828 A JP 2001-282549 A JP 2007-257384 A JP-A-8-147156

  An object of the present invention is to provide a job network automatic generation method for converting a JCL file described in a JCL language used in a mainframe into a simple job network flow.

  A job network automatic generation method according to a first aspect of the present invention includes a JCL file input unit, a job network generation unit, and a job network output unit. The JCL file input unit inputs a JCL file. The job network generation unit automatically generates a flow link based on the JCL description of the JCL file. The job network output unit converts the flow link into a job network flow and outputs it.

  The program according to the second aspect of the present invention causes a computer to function as JCL file input means, job network generation means, and job network output means. The JCL file input means inputs a JCL file. The job network generation means automatically generates a flow link based on the JCL description of the JCL file. The job network output means converts the flow link into a job network flow and outputs it.

  According to the present invention, it is possible to provide a job network automatic generation method for converting a JCL file described in the JCL language used in the mainframe into a simple job network flow.

FIG. 1-1 is a first diagram illustrating a job network flow representing the same order relationship. FIG. 1-2 is a second diagram illustrating a job network flow representing the same order relationship. FIG. 1-3 is a third diagram showing a job network flow representing the same order relationship. FIG. 1-4 is a fourth diagram illustrating a job network flow representing the same order relationship. FIG. 2 is a diagram showing an overall configuration of the job network automatic generation method. FIG. 3 is a configuration diagram of the JCL → intermediate file conversion unit. FIG. 4 is a configuration diagram of the job network generation unit. FIG. 5 is a functional explanatory diagram of the script file dividing unit. FIG. 6 is a functional explanatory diagram of the flow link creation unit. FIG. 7 is a first explanatory diagram of the flow link creation method. FIG. 8 is a second explanatory diagram of the flow link creation method. FIG. 9 is a third explanatory diagram of the flow link creation method. FIG. 10 is a fourth explanatory diagram of the flow link creation method. FIG. 11 is a fifth explanatory diagram of the flow link creation method. FIG. 12 is a sixth explanatory diagram of the flow link creation method. FIG. 13 is a flowchart for explaining processing by the JCL → intermediate file conversion unit. FIG. 14 is a first flowchart illustrating the flow link creation method. FIG. 15 is a second flowchart illustrating the flow link creation method. FIG. 16 is a third flowchart illustrating the flow link creation method. FIG. 17 is a fourth flowchart illustrating the flow link creation method. FIG. 18 is a flowchart illustrating the output processing of the job network output unit.

  One of the best modes for carrying out the present invention will be described in detail with reference to the drawings. FIG. 2 shows the overall configuration of the job network automatic generation method according to one embodiment. In FIG. 2, the job network automatic generation method 10 includes a JCL file input unit 11, a JCL → intermediate file conversion unit 12, a job network generation unit 13, and a job network output unit 14. The JCL file input unit 11 reads a JCL file to be converted from the storage device A1. The JCL → intermediate file conversion unit 12 analyzes the JCL file and converts the JCL file into an intermediate file described in a shell language that can be interpreted by a general-purpose UNIX (registered trademark) OS. The job network generation unit 13 analyzes the JCL description of the intermediate file, creates a flow link, and further divides the intermediate file into job scripts. The job network output unit 14 converts and outputs the created flow link and script into a job network definition file. The job network automatic generation method 10 in FIG. 2 inputs a JCL file group from the storage device A1, and outputs a job network definition group to the storage device B1.

  FIG. 3 shows a configuration diagram of the JCL → intermediate file conversion unit 12. In FIG. 3, the JCL → intermediate file conversion unit 12 includes an intermediate file generation unit 32, an intermediate file output unit 34, and a conversion dictionary 36. The intermediate file generation unit 32 converts the JCL file extracted from the storage device A1 by the JCL file input unit 11 into an intermediate file. The intermediate file output unit 34 outputs the generated intermediate file. The conversion dictionary 36 is prepared in advance, and the intermediate file generation unit 32 performs conversion processing while referring to the conversion dictionary 36. The conversion dictionary 36 stores a correspondence table from JCL to shell script.

  FIG. 4 shows a configuration diagram of the job network generation unit 13. In FIG. 4, the job network generation unit 13 includes a JCL analysis unit 41, a script file division unit 42, a flow link creation unit 43, a flow link diagnosis unit 44, and a flow link optimization unit 45. . The JCL analysis unit 41 scans and analyzes the JCL description in the intermediate file. The script file dividing unit 42 receives the actual processing line of the JCL description from the JCL analyzing unit 41 and stores it in the script file of the storage device B1. The flow link creation unit 43 receives the JCL instruction line of the JCL description from the JCL analysis unit 41, and adds the flow components to the flow link of the main memory MEM1 according to the type. The flow link diagnosis unit 44 diagnoses whether or not the flow link created by the flow link creation unit 43 has a correct structure. The flow link optimization unit 45 adjusts the structure of the flow link and optimizes the flow link.

  FIG. 5 is a functional explanatory diagram of the script file dividing unit 42. In FIG. 5, the JCL analysis unit 41 linearly scans the input JCL description, and determines whether the scan line is a JCL instruction line (control line) or other actual processing line. If it is determined that the line is a JCL instruction line, the JCL analysis unit 41 passes the JCL instruction line to the flow link creation unit 43. On the other hand, when it is determined that it is an actual processing line, the JCL analysis unit 41 passes the actual processing line to the script file division unit 42.

  The script file dividing unit 42 divides and stores an actual processing line in which actual processing other than the JCL instruction line is described in a plurality of script files created in the storage device B1. Each script file is opened (OPEN) and closed (CLOSE) by the flow link creation unit 43 as a script file associated with any trunk or branch by the appearance of a JCL command line.

  FIG. 6 is a functional explanatory diagram of the flow link creation unit 43. As illustrated, the flow link creation unit 43 includes a trunk management unit 52, a branch management unit 54, and a waiting management unit 56, and creates a flow link using the main memory MEM1. The flow link creation unit 43 receives the JCL command line from the JCL analysis unit 41, and distributes the received JCL command line to the trunk management unit 52, the branch management unit 54, and the waiting management unit 56.

The flow link creation unit 43 interprets the control flow of the JCL description as follows.
1. The line between the \ SUBJOB instruction and the \ ENDSUBJOB instruction is recognized as a “branch”.
2.1. All lines such as JCL instruction lines, \ JOB instruction and \ ENDJOB instruction are recognized as “stem”.
3. The \ WAITSUB instruction is considered as a waiting point (convergence point) for branch processing by \ SUBJOB.
4). JCL is considered to be a process in which zero or more “branch” processes are derived from the basic “trunk” process and converge by waiting.

The flow link creation unit 43 creates a flow link using the following components.
・ Start (START) parts ・ End (END) parts ・ Unit job parts ・ Parallel branch parts ・ Event transmission parts ・ Event reception parts

  A flow link creation method by the flow link creation unit 43 will be described with reference to FIGS. First, the flow link creation unit 43 stores the start part in the flow link developed in the main memory MEM1. Next, when the JCL analysis unit 41 sends ¥ JOB Daily (¥ STEP JOB1) of the JCL instruction line to the flow link creation unit 43, the trunk management unit 52 updates the flow and adds the unit job part trunk 1. . Subsequently, when the JCL analysis unit 41 sends \ SUBJOB SUB1 (\ STEP JOB2) of the JCL instruction line to the flow link creation unit 43, the branch management unit 54 updates the flow, and the parallel branch parts 1 and 1 Add Then, when the JCL analysis unit 41 sends ¥ JOB3 (¥ STEP JOB3) of the JCL command line to the flow link creation unit 43, the trunk management unit 54 holds the trunk 1 of the current unit job part, and a waiting point. A trunk 2 is generated as a new unit job part to be associated with.

  FIG. 7 is an operation explanatory diagram when the flow link creation unit 43 receives ¥ SUBJOB SUB2 (¥ STEP JOB4) in the JCL instruction line. In FIG. 7, upon receiving ¥ SUBJOB SUB2, the branch management unit 54 attempts to newly add parallel 2 and branch 2 of the parallel branch parts in order to update the flow. As shown in the drawing, the branch management unit 54 manages branches, arranges parallel 2 and branch 2 of parallel branch parts in parallel with parallel 1 and branch 1 of parallel branch parts, and updates the flow. That is, the branch management unit 54 adds a parallel branch component representing parallel execution to the link flow in order to execute the processing contents up to the \ SUBJOB instruction and the ENDSUBJOB instruction as the branches in parallel. At this time, the next branch is newly created in the branch path, and the convergence is managed. Further, the script file is associated with the branch n, and all the lines from \ SUBJOB to \ ENDSUBJOB are associated with the branch n and stored.

  The trunk management unit 52 manages the trunk portion. The trunk management unit 52 opens and closes the script file so as to record all the actual processing lines in the same script file until the \ SUBJOB instruction or the \ WAITSUB instruction is generated. At this time, the script file dividing unit 42 records all the actual processing lines in the same script file associated with the trunk.

  FIG. 8 is an operation explanatory diagram when the flow link creation unit 43 receives ¥ WAITSUB of the JCL command line. The trunk management unit 52 has already created the trunk 2 to be associated with the waiting point at the stage before receiving \ WAITSUB. As shown in the figure, when the flow link creation unit 43 receives ¥ WAITSUB, the waiting management unit 56 adds waiting 1 of the waiting part (also referred to as a focus (FOCUS) part) after the latest trunk 2. The trunk management unit 52 closes the script file associated with the trunk 1 and opens the script file associated with the trunk 2.

  9 to 11 are diagrams for explaining a branch convergence method. When the waiting part 1 of the waiting part is added, the flow link creation unit 43 converges the branches of the flow link. As illustrated in FIG. 9, the branch management unit 54 requests the waiting management unit 56 to search for a branch to be converged described as an argument of the \ WAITSUB line. First, the branch 1 and the waiting 1 are joined. In addition, the trunk management unit 52 joins the trunk 2 behind the parallel 2. As a result, the flow link is updated as shown in FIG.

  Subsequently, the branch management unit 54 tries to converge the branch 2. The queuing management unit 56 newly generates the queuing part 2 of the queuing part, and combines this queuing 2 behind the queuing 1. The branch management unit 54 combines the branch 2 and the waiting 2. In this way, the process related to reading \ WAITSUB is completed. When the JCL analysis unit 41 further advances scanning, the flow link creation unit 43 further combines the unit job component trunk 3 and the end component. As a result, the flow link is updated as shown in FIG.

  When the scanning of the JCL description is completed, the flow link diagnosis unit 44 uses the branch management unit 54 to diagnose branch parts. If there is an unconverged branch, a warning to that effect is given. Next, the flow link optimization 45 uses the branch optimization unit 58 to optimize the flow link according to the flow method of the job management software. As shown in FIG. 11, the branch optimization unit 58 sorts the parallel branch parts and eliminates the intersection of the branches. FIG. 12 is a diagram illustrating the input / output relationship of the job network generation unit 13. As illustrated, the job network generation unit 13 inputs a JCL description of an intermediate file, and outputs a flow link and a script file group.

  FIG. 13 is a flowchart for explaining processing of the JCL → intermediate file conversion unit 12. In FIG. 13, the JCL → intermediate file conversion unit 12 opens the JCL file (step 21-1) and reads the JCL description line by line (step 21-2). The intermediate file generation unit 32 repeats the conversion process until an EOF (End Of File) of the JCL file (step 21-3). While the ECL of the JCL file has not been reached (step 21-3: YES), the intermediate file generation unit 32 searches the conversion dictionary 36 (step 21-4), and the read line is set as the conversion target line. 36 is checked (step 21-5). If it is a conversion target line, conversion is performed and the search result of the conversion dictionary 36 is output to an intermediate file (step 21-8). If it is not a conversion target line, the read line is output as it is (step 21-7). When the EOF of the JCL file is reached (step 21-3: NO), the intermediate file is closed (step 21-6). The intermediate file output unit 34 transmits the created intermediate file as a conversion result to the job network generation unit 13. By this conversion processing, the JCL file is converted into an intermediate file described in the shell language. Further, the JCL control command line necessary for the subsequent conversion process is left as a comment line.

14 to 17 are flowcharts for explaining a method of creating a flow link by the job network generation unit 13. The job network generation unit 13 in the present embodiment generates a flow link for the following JCL command line.
¥ JOB Start of JCL description ¥ ENDJOB End of JCL description ¥ SUBJOB Start of parallel execution processing description ¥ ENDSUBSUB Job end of parallel execution processing description ¥ WAITSUB Wait for end of specified SUBJOB ¥ RUN Run the specified JCL file

  As shown in FIG. 14, in the job network generation unit 13, when the JCL analysis unit 41 opens the intermediate file received from the JCL → intermediate file conversion unit 12 (step 22-1), the intermediate file is read line by line (step 22). -2). The JCL analysis unit 41 repeats the reading of the intermediate file until EOF (step 22-3). If the EOF of the intermediate file is reached (step 22-3: NO), the analysis of the intermediate file is terminated or an error is output.

  Until the EOF of the intermediate file is reached (step 22-3: YES), it is determined whether the read line is an actual processing line or a JCL instruction line (step 22-4). If it is an actual processing line (step 22-4: No), the next line is read. If it is a JCL command line (step 22-4: YES), the job network flow is opened (step 22-5), and the process proceeds to the detailed analysis process of FIG.

  In the detailed analysis process of FIG. 15, the flow link creation unit 43 first adds a start part to the base (BASE) path as an initialization process, then adds a unit job part, and starts creating a flow link. To do. Further, a new script file associated with the unit job part is created in the storage device B1 and opened (step 23-1). If the target line is not a JCL command line, the script file dividing unit 42 stores the line in the opened script file.

  The JCL analysis unit 41 reads the intermediate file line by line (step 23-2), and repeats this reading until EOF (step 23-3). If the read target line is a \ SUBJOB line (step 23-4), the flow link creation unit 43 closes the unit job part of the base path (step 23-9) and sets the parallel branch part to the base path. It adds (step 23-10). Furthermore, a new unit job part is created in the parallel branch path (step 23-11). The script file dividing unit 42 stores all lines up to ¥ ENDSUBJOB in the script file associated with the unit job component. Thereafter, the process proceeds to parallel branch path processing (step 23-12).

  If the read target line is a \ WAITSUB line and there is no JOBNAME argument (step 23-5), the flow link creation unit 43 closes the unit job part of the base path (step 23-13). . Next, a focus (FOCUS) part is added to the base path, a branch to be converged described as an argument of the \ WAITSUB line is searched, and all parallel branch parts are converged by this focus part (step) 23-14). Subsequently, the next unit job part is newly created in the base path and opened (step 23-15).

  If the read target line is the \ WAITSUB line and the argument is JOBNAME (step 23-6), the flow link creation unit 43 closes the unit job part of the base path (step 23-16). ). Next, a focus part is added to the base path, a path of JOBNAME described as an argument in the \ WAITSUB line is searched, and these parallel branch parts are converged by this focus part (step 23-17). Subsequently, the next unit job part is newly created in the base path and opened (step 23-18).

  If the read target line is a \ ENDJOB line (step 23-7), the flow link creation unit 43 ends the creation of the flow link and closes the job network flow (step 23-19).

  If the read target line is a \ RUN line and the parameter value is HOLD = N (step 23-8), the flow link creation unit 43 proceeds to the \ RUN process (step 23-20). The parameter value HOLD = N means that the JOBNAME is executed (¥ RUN) after waiting for the argument JOBNAME to be validated N times (¥ ACTIVATE).

  The flow link creation unit 43 determines that the read target line is not a \ SUBJOB line, a \ WAITSUB line, a \ ENDJOB line, or a parameter value of HOLD = N and is not a \ RUN line. Determines that the read line is not an identifier line (step 23-21). The script file dividing unit 42 adds this line to the script file associated with the opened unit job component (step 23-22).

  FIG. 16 is a flowchart for explaining parallel branch path processing. In the parallel branch path processing of FIG. 16, the flow link creation unit 43 first opens a unit job part in the parallel branch path (step 24-1). Next, the JCL analysis unit 41 reads the intermediate file line by line (step 24-2), and repeats this reading until EOF (step 24-3). The flow link creation unit 43 determines whether the read target line is “¥ RUN HOLD = ON” or “¥ ACTIVATE” (step 24-4). If it is either, the process proceeds to the \ RUN process (step 24-5). If neither is found, it is determined whether or not it is “¥ ENDSUBJOB” (step 24-6). If “¥ ENDSUBJOB”, the unit job part is closed (step 24-7), the parallel branch path is closed (step 24-8), and the parallel branch path processing is terminated. If it is not “¥ ENDSUBJOB”, it is determined that the line is an actual process line (step 24-9), and the line is added to the script file associated with the unit job component (step 24-10).

  FIG. 17 is a flowchart for explaining the RUN process. In FIG. 17, when “¥ RUN HOLD = ON”, the unit job part is closed (step 25-1), and N event reception parts are created (step 25-2). Subsequently, a sub job network “JOBNAME” is created in the flow link (step 25-3), and the next unit job is opened (step 25-4). On the other hand, if it is not “¥ RUN HOLD = ON”, the unit job part is closed (step 25-5), and an event transmission part is created (step 25-6). Subsequently, the next unit job part is opened (step 25-7).

  When the scanning of the JCL description is completed, the queuing management unit 56 scans the branch (BRANCH) component, and warns that if there is an unconverged branch component. Next, the flow link optimization unit 45 sorts the branch parts according to the index of the focus parts.

  FIG. 18 is a flowchart for explaining output processing of the job network output unit 14. The job network output unit 14 converts a flow link or the like into a job network definition corresponding to the type of job management software in accordance with the flowchart of FIG. In FIG. 18, first, the job network output unit 14 determines whether or not the link structure of the flow link is normal (step 26-1). If it is not normal, an error is output. If it is normal, it is determined whether or not it is a structure that can be drawn as a job network or a structure that can output a file (step 26-2). If it is such a structure, the link structure is directly rendered as a job network flow or outputted as a file (step 26-3). If it is not such a structure, it is determined whether or not it can be repaired (step 26-4). If it can be repaired, the link structure is corrected (step 26-5).

  As described above, according to the present embodiment, the following effects can be obtained. The first effect is that the cost of migrating the job environment from mainframe to open system is greatly reduced. The reason is that, according to the present embodiment, automatic conversion of JCL → job network becomes possible, thereby minimizing the man-hour and risk of job migration that has been performed manually.

  The second effect is that the maintainability after shifting to the job environment is greatly improved. The reason for this is that the job network flow automatically generated by this embodiment is a simple flow with high maintainability, and job monitoring and job correction work after migration is more efficient than the JCL file. This makes it extremely easy.

  A third effect is that the maintainability of job operation in general job management software can be improved. The reason is that the algorithm according to the present embodiment, which converts a complex flow into a flow having a simple shape, is applied to the flow optimization of general job management software, so that the flow with high maintainability is applied. This is because it is possible to convert to

10 Job Network Automatic Generation Method 11 JCL File Input Unit 12 JCL → Intermediate File Conversion Unit 13 Job Network Generation Unit 14 Job Network Output Unit 32 Intermediate File Generation Unit 34 Intermediate File Output Unit 36 Conversion Dictionary 41 JCL Analysis Unit 42 Script File Division Unit 43 Flow link creation unit 44 Flow link diagnostic unit 45 Flow link optimization unit 52 Trunk management unit 54 Branch management unit 56 Wait management unit 58 Branch optimization unit

Claims (7)

JCL file input means for inputting a JCL (Job Control Language) file;
The JCL description of the JCL file includes a JCL actual processing line and a JCL instruction line,
Job network generation means for automatically generating a flow link indicating a job control flow based on the JCL command line;
Job network output means for converting and outputting the flow link into a job network definition of a job network flow that is a flowchart of a job execution sequence;
The job network generation means includes
JCL analysis means that scans the JCL description and distributes it into a JCL actual processing line and a JCL instruction line;
With flow link creation means,
The flow link creating means includes:
Branch management means for receiving the JCL instruction line, receiving a parallel execution process instruction line and a parallel execution process end instruction line in the JCL instruction line, and managing the line between them as a branch;
A trunk management means for receiving a JCL instruction line other than the parallel execution process instruction line and the parallel execution process end instruction line in the JCL instruction line, and managing a line other than the branch as a trunk;
The trunk management means includes
A unit job part (stem part) indicating an execution position of a JCL command line managed as one trunk is associated with one trunk, and the trunk part is added to the flow link being created,
The branch management means includes
A set of parallel parts and branch parts is associated with one branch, and the parallel parts and the branch parts are added to the flow link being created, where the parallel parts are managed as the branches. A job network automatic generation method that indicates a start position of processing of a JCL actual processing line, and the branch component indicates an execution position of a JCL actual processing line managed as the branch. The flow link creating means includes:
Among the JCL instruction lines, a parallel execution process synchronization instruction line is received, and the branch to be converged is managed by searching for the branch described as an argument of the parallel execution process synchronization instruction line. Further comprising waiting management means,
The waiting management means includes:
When a parallel execution process synchronization instruction line is received, a waiting part representing the end of execution of the JCL actual processing line corresponding to the branch part is added to the branch part corresponding to the parallel execution process synchronization instruction line in the flow link being created. The job network automatic generation method according to claim 1, wherein the flow links are updated by combining. The job network generation means includes
Referring to the index of the waiting part corresponding to the parallel part of the flow link, sort the processing context between the parallel parts, and calculate the intersection of the parallel branch paths caused by the combination of the branch part and the waiting part. The job network automatic generation method according to claim 2, further comprising a flow link optimization unit for canceling. The job network generation means includes
4. The job network automatic generation method according to claim 3, further comprising script file dividing means for dividing the JCL actual processing line into a plurality of script files and storing them. The flow link creating means includes:
Corresponding the trunk and the unit job part, or the branch and the branch part, and open a script file associated with the unit job part or branch part,
The script file dividing means includes:
The job network automatic generation method according to claim 4, wherein a JCL actual processing line corresponding to the unit job part or the branch part is stored in a script file opened by the flow link creation unit. The flow link creating means includes:
When the script file dividing means finishes storing the JCL actual processing line corresponding to the script file when the script file is open, the script file is closed,
6. The job network automatic generation method according to claim 5, wherein when there is another unit job part or the branch part, a script file associated with the unit job part or the branch part is opened. A program for causing a computer to function as a JCL file input unit, a job network generation unit, and a job network output unit included in the automatic job network generation method according to claim 1.

Images