JPWO2018061219A1 - Job scheduling system, job scheduling method, and job scheduling apparatus - Google Patents

Abstract

A job scheduling system, a job scheduling method, and a job scheduling apparatus that can be used to redesign a plurality of jobs are provided. SOLUTION: For each of a plurality of jobs, history information including time information indicating time when a plurality of jobs are started and access information indicating that a file is referred to or updated in each of a plurality of jobs Identify the context, and identify the job that refers to the file among the jobs between the job of the starting point that has updated the file and the job of the end point that indicates the latest job that has updated the file The degree of influence is set according to the identified job, and a generation unit for generating a schedule proposal indicating the degree of influence and an output unit for outputting the schedule proposal are provided.

Description

  The present invention relates to a job scheduling system, a job scheduling method, and a job scheduling apparatus, and is suitably applied to, for example, a job scheduling system used for redesigning a schedule of a plurality of jobs, a job scheduling method, and a job scheduling apparatus. .

  Generally, in a system represented by a financial system etc., online processing and batch processing are performed. In particular, batch processing is often performed after the online work in the daytime is over, since it handles the tabulation processing of data handled in the online process in the daytime. Therefore, it may be called night batch processing. In nighttime batch processing, a large number of jobs selected according to conditions are executed out of a huge number of jobs such as tens of thousands or more. Then, a device has been devised to complete the batch work (second process) by the start of the online work (first process) (see Patent Document 1).

  In recent years, technology to select jobs that can operate under current computer resource utilization conditions by referring to past job operation results (see Patent Document 2), and find the longest processing time of each job from operation results, A technology for determining the priority of a job on the basis is disclosed (see Patent Document 3).

JP, 2015-170189, A JP, 2015-64664, A JP, 2014-119860, A

  By the way, since the structure of batch processing is also changed along with the change of business form, the schedule of batch processing that was optimized at the beginning is largely different from the original because a large number of jobs are added, etc. The result may be a complex system that can not be analyzed at the end of the statement. However, with the technology described in the above-mentioned patent documents, although it is possible to determine and execute the execution order of jobs in real time, there is a problem that it is not possible to redesign a complicated system job.

  The present invention has been made in consideration of the above points, and an object thereof is to propose a job scheduling system, a job scheduling method, and a job scheduling apparatus which can be used for redesigning a plurality of jobs.

  In order to solve such problems, the present invention is a job scheduling system used for redesigning a schedule of a plurality of jobs, wherein time information indicating times when the plurality of jobs are started, each of the plurality of jobs For each of the plurality of jobs, from the history information including access information indicating that the file has been referenced or updated, the context of the job is specified for each of the plurality of jobs, and the subsequent job is the starting job that updated the file and the file Of the jobs between the end job indicating the most recent job that has further updated the file and the job referring to the file among the jobs, and setting the degree of influence according to the specified job, the schedule proposal indicating the degree of influence , And an output unit for outputting the schedule proposal.

  Further, according to the present invention, there is provided a job scheduling method executed by a system used for redesigning a schedule of a plurality of jobs, wherein time information indicating times when the plurality of jobs are started, a file for each of the plurality of jobs For each of the plurality of jobs, the context of the job is specified from the history information including the access information indicating that the file is referred to or updated, and the job and the file that are the subsequent jobs and the file is updated Further, a job referring to the file is specified among the jobs between the job and the end point indicating the latest job updated, the influence degree is set according to the specified job, and the schedule proposal indicating the influence degree A first step of generating and a second step of outputting the proposed schedule are provided.

  Further, in the present invention, it is a job scheduling apparatus used for redesigning a schedule of a plurality of jobs, and time information indicating time when the plurality of jobs are started, and a file is referred or updated in each of the plurality of jobs. For each of the plurality of jobs, from the history information including the access information indicating that it has been done, the context of the job is specified, and it is the succeeding job and the latest job of the starting point that updated the file and the latest updated file. The job that identifies the job that refers to the file among the jobs between the job and the end job that indicates the job, sets the degree of influence according to the specified job, and generates a schedule proposal that indicates the degree of influence And an output unit for outputting the schedule proposal.

  According to the job scheduling system, the job scheduling method, and the job scheduling apparatus which can be used for redesigning a plurality of jobs of the present invention, there is provided a schedule proposal indicating the degree of influence corresponding to the subsequent job for each of the plurality of jobs. As it is generated and output, multiple jobs can be redesigned based on the degree of influence.

  According to the present invention, it is possible to realize a job scheduling system, a job scheduling method, and a job scheduling apparatus that can be used to redesign a plurality of jobs.

FIG. 1 is a diagram showing a schematic configuration of a job scheduling system. It is a figure which shows the data structure of the data registered into execution log | history DB. It is a figure which shows the data structure of the data registered into execution log | history DB. It is a figure which shows the data structure of the data registered into execution log | history DB. It is a figure which shows an example of the data registered into execution history DB. It is a figure which shows the data structure of the job information storage table which stores job information. It is a figure which shows the data structure of the file information storage table which stores file information. It is a figure which shows an example of a management table queue. It is a figure which shows the flowchart which concerns on a schedule analysis process. It is a figure which shows the flowchart which concerns on a schedule proposal preparation process. It is a figure which shows the flowchart which concerns on a schedule proposal display process. It is a figure which shows the example of a display of a schedule proposal.

  An embodiment of the present invention will now be described in detail with reference to the drawings.

(1) Configuration of Job Scheduling System In FIG. 1, 1 generally shows a job scheduling system according to the present embodiment. The job scheduling system 1 is a system that presents information used when designing (redesigning) job submission procedures, and includes a first computer 10, a second computer 20, a file for batch 30, a first The execution history DB 40 of the computer 10, the schedule proposal DB 50, and the schedule proposal output device 60 are provided.

  The first computer 10 is a computer or the like, and includes a CPU 11 and a memory 12. The CPU 11 reads out various programs to the memory 12 and executes them to execute a batch program 13, a job schedule management unit 14, an execution history management unit 15. And the OS 16 is realized. In the first computer 10, the job schedule management unit 14 manages the execution order of the batch program 13 (each job) to execute the batch program 13, and the execution history management unit 15 executes the execution result of the batch program 13 such as a hard disk It registers in execution history DB40 provided in a memory storage.

  The second computer 20 is a computer or the like, and includes a CPU 21 and a memory 22. The CPU 21 reads out various programs to the memory 22 and executes the programs to generate a schedule analyzing unit 23, a schedule proposal creating unit 24, and a schedule proposal display unit 25. , And the OS 26 are realized. In the second computer 20, the schedule analysis unit 23 reads out the execution history (record) from the execution history DB 40, analyzes the schedule (such as the anteroposterior relationship) of the job, and the schedule proposal creating unit 24 analyzes the schedule for each job. The schedule proposal is created by registering each job in order to determine the degree of influence on the subsequent jobs, and is registered in the schedule proposal DB 50, and the schedule proposal display unit 25 reads the schedule proposal from the schedule proposal DB 50, and the schedule proposal Output to the output device 60.

(2) Data Configuration of Execution History and Data Example of Execution History FIGS. 2 to 4 are diagrams showing the data structure of data (records) registered in the execution history DB 40. FIG. FIG. 2 shows the data structure of the job start information 41. The job start information 41 includes a record identifier "01", a job name, and a job start time. FIG. 3 shows the data structure of the job end information 42. The job end information 42 includes a record identifier "02", a job name, and a job end time. FIG. 4 shows the data structure of the file close information 43. In the file close information 43, record identifier "03", job name, file name, reference / update identifier indicating reference identifier or update identifier, and reference count or update The number of references / updates that indicates the number is included.

  FIG. 5 is a diagram showing an example of data (record) registered in the execution history DB 40. As shown in FIG. In the execution history DB 40, job start information 41 is registered when the job starts, job end information 42 when the job ends, and file close information 43 when the job accesses (refers to / updates) the file. Be accumulated. For example, execution of job 1 is started to “××: × 1: ××”, and in job 1, file 11 is referenced “×” times, file 12 is updated “×” times, and job 1 execution is When ending in “XX: × 6: XX”, records 44 to 47 are registered as shown in FIG. 5. The record 44 is an example of the job start information 41, the record 45 and the record 46 are an example of the file close information 43, and the record 47 is an example of the job end information 42.

(3) Example of job schedule analysis (example of management table queue)
FIG. 6 and FIG. 7 are diagrams showing the data structure of tables constituting the management table queue 90 shown in FIG. FIG. 6 shows the data structure of the job information table 70 storing job information, and FIG. 7 shows the data structure of the file information table 80 storing file information.

  As shown in FIG. 6, the job information table 70 includes, as job information, a forward pointer 71, a backward pointer 72, a job name 73, a job start time 74, a job end time 75, an access file queue top pointer 76, and an access file. The information of queue end pointer 77, impact evaluation value 78, and priority 79 is included.

  As shown in FIG. 7, the file information table 80 includes, as file information, a forward pointer 81, a backward pointer 82, a file name 83, a reference / update identifier 84 indicating a reference identifier or an update identifier, the number of accesses 85, and reference The information of the file generation flag 86 is included.

  For example, as shown in FIG. 8, the management table queue 90 is referred to or updated in each job of the plurality of job information tables 70 (70a, 70b, 70c,...) And the plurality of job information tables 70. Or one or more file information tables 80 (80-1a, 80-2a, 80-1b, 80-2b, 80-1c, 80-2c,...) Corresponding to a plurality of files. Be done.

  For each job information table 70, the forward / backward relationship of the job is specified (identified) by the forward pointer 71 (pointers 71a, 71b, 71c,...) And the backward pointer 72 (pointers 72b, 72c,. . Also, for each file information table 80, the access file queue head pointer 76 (pointers 76a, 76b, 76c, ...), the access file queue final pointer 77 (pointers 77a, 77b, 77c, ...), the forward pointer 81 (Pointers 81-2a, 81-2b, 81-2c,...) And backward pointer 82 (Pointers 82-2a, 82-2b, 82-2c,. Files to be accessed by the job are identified. The number of job information tables 70 and the number of file information tables 80 shown in FIG. 8 are merely examples, and the number is not limited to these numbers.

(4) Schedule Analysis FIG. 9 is a diagram showing a flowchart related to schedule analysis processing. As shown in FIG. 9, the schedule analysis unit 23 reads one record from the execution history DB 40 (S1), and determines whether the read record is "EOF (End Of File)", that is, the last record (S2). ). If the schedule analysis unit 23 determines that it is "EOF", it ends the process, and if it determines that it is not "EOF", it determines whether the record identifier of the read record is "01", that is, the job starts It is determined whether the record is a record indicating (S3). If the schedule analysis unit 23 determines that the record identifier is "01", the process proceeds to S4, and if it is determined that the record identifier is not "01", the process proceeds to S8.

  In S4, the schedule analysis unit 23 prepares a table (job information table 70) for storing job information. Subsequently, the schedule analysis unit 23 registers the job name and the job start time of the read record in the job name 73 and the job start time 74 of the prepared job information table 70 (S5), and the job information table in which the information is registered. 70 is connected to the management table queue 90 (S6), and the process moves to S7.

  More specifically, in S6, the schedule analysis unit 23 registers pointers in each of the forward pointer 71 and the backward pointer 72 of the job information table 70. For example, when preparing the job information table 70 b shown in FIG. 8, the schedule analysis unit 23 registers the pointer 71 b in the forward pointer 71 and the pointer 72 b in the backward pointer 72.

  In S7, the schedule analysis unit 23 reads the next record from the execution history DB 40, and shifts the processing to S2.

  In S8, the schedule analysis unit 23 determines whether the record identifier of the read out record is “02”, that is, whether the record indicates the end of the job. If the schedule analysis unit 23 determines that the record identifier is "02", the process proceeds to S13, and if it is determined that the record identifier is not "02", the process proceeds to S9.

  In S9, the schedule analysis unit 23 searches the management table queue 90 for an entry (job information table 70) in which the job names are equal and the job end time is not registered. Subsequently, the schedule analysis unit 23 prepares a table (file information table 80) for storing file information (S10), and sets the file name 83, the reference / update identifier 84, and the access count 85 of the prepared file information table 80. The file name of the read record, the reference / update identifier, and the reference / update count are registered (S11), the file information table 80 in which the information is registered is connected to the management table queue 90 (S12), and the process is transferred to S7.

  More specifically, in S12, the schedule analysis unit 23 updates the access file queue top pointer 76 and the access file queue final pointer 77 of the retrieved job information table 70, and the forward pointer 81 of the prepared file information table 80 and A pointer is registered in each of the backward pointers 82. For example, when preparing the file information table 80-1b shown in FIG. 8, the schedule analysis unit 23 updates each of the access file queue top pointer 76 and the access file queue last pointer 77 to the pointer 76b, and the forward pointer 81 and back A null pointer is registered in each of the word pointers 82. Further, for example, when the schedule analysis unit 23 prepares the file information table 80-2b shown in FIG. 8, the access file queue final pointer 77 is updated to the pointer 77b, and the forward pointer 81 is changed to the pointer 81-2b and backward pointer. The pointer 82-2b is registered at 82.

  In S13, the schedule analysis unit 23 searches the management table queue 90 for an entry (job information table 70) in which the job names are equal and the job end time is not registered. Subsequently, the schedule analysis unit 23 registers the job end time of the read record in the job end time 75 of the retrieved job information table 70 (S14), and shifts the processing to S7.

  As described above, by performing the schedule analysis process, the management table queue 90 indicating the analysis result of the job (the job's context, file access, etc.) is completed.

(5) Schedule Proposal Creation FIG. 10 is a diagram showing a flowchart related to schedule proposal creation processing. As shown in FIG. 10, the schedule proposal creating unit 24 sets the top entry (job information table 70) of the management table queue 90 as a verification target (S21), and moves the process to S22.

  In S22, the schedule proposal creating unit 24 sets the impact evaluation value 78 of the job information table 70 to be verified to "0". Subsequently, the schedule proposal creating unit 24 sets the top entry (file information table 80) of the file information managed by the job information table 70 to be verified as the verification target (S23), and moves the process to S24.

  In step S24, the schedule proposal creating unit 24 determines whether the reference / update identifier of the file information table 80 to be verified is "update". If it is "update", the process proceeds to step S25. If not "update", the process moves to S32.

  In S25, the schedule proposal creating unit 24 sets the job information table 70 and the file information table 80 currently targeted as the starting point of the verification target. Subsequently, the schedule proposal creating unit 24 sequentially searches the job information table 70 and the file information table 80 following the starting point table, and first, the file name is equal and the reference / update identifier is "update". The file information table 80 satisfying the condition and the job information table 70 related to the file information table 80 are extracted (S26), and it is determined whether a table satisfying the condition is extracted (S27). If the schedule proposal creating unit 24 determines that the job information table 70 has been extracted as the job information table 70 of the end point (S 28), and if it is determined that the job information table 70 has not been extracted, the last entry of the management table queue 90 (job information The table 70) is made the job information table 70 of the end point (S29), and the process is shifted to S30.

  In S30, the schedule proposal creating unit 24 sets (1) the file information table of the starting point from the file information table 80 pointed to by the job information table 70 on the queue of the job information table 70 of the starting point and the job information table 70 of the ending point. Set the reference file creation flag 86 of the file information table 80 which satisfies these conditions to "ON" which is equal to the file name described in 80, (2) the reference / update identifier of the file information table 80 is "reference". Do.

  Subsequently, for each job information table 70, the schedule proposal creating unit 24 selects a file information table in which the reference file generation flag 86 is "ON" from the group of file information tables 80 connected to the queue pointed to by the table. 80 is counted and stored in the impact degree evaluation value 78 of the job information table 70 to be verified (S31), and the process is shifted to S32.

  In S32, the schedule proposal creating unit 24 determines whether the file information table 80 to be verified is the last of the queue, and if it is determined not to be the last, sets the next file information table 80 as a verification target (S33) , S24, if it is determined to be the last, the process moves to S34.

  In S34, the schedule proposal creating unit 24 determines whether the job information table 70 to be verified is the last of the queue, and if it is determined not to be the last, sets the next job information table 70 as a verification target (S35) , And when it is determined that the process is not the final one, the process proceeds to S36. Note that, at S35, the schedule proposal creating unit 24 sets all reference file generation flags 86 to "OFF".

  In S36, the schedule proposal creating unit 24 lists up the job information table 70 in which the influence degree evaluation value 78 is stored. Subsequently, the schedule proposal creating unit 24 stores the order numbers (priority) in the priority order 79 of the job information table 70 in descending order of the impact degree evaluation value from the job information table 70 listed up (S37). finish. If there is a job information table 70 with the same impact degree evaluation value 78, the priority 79 of the job information table 70 with the earlier job start time 74 is set higher.

  As described above, for each of the plurality of jobs, the subsequent job is added based on the job context and the file access relationship, and a schedule proposal is generated in which the priority (impact evaluation value) is indicated.

(6) Schedule Proposal Display FIG. 11 is a diagram showing a flowchart related to the schedule proposal display process. As shown in FIG. 11, the schedule proposal display unit 25 causes the schedule proposal output device 60 to display the job information with the first priority in the job information table 70 (S41), and the job information table 70 of the next priority ( Job is searched, the job information of the job is displayed on the schedule proposal output device 60 (S42), and whether or not the job is the final entry, that is, whether all job information of the job to be displayed is displayed If it is determined that the entry is not the final entry, the process proceeds to S42. If it is determined that the entry is the final entry, the process ends.

  According to the above-described process, the proposed schedule is output to the proposed schedule output device 60.

(7) Display Example of Schedule Proposal FIG. 12 shows a display example of the schedule proposal. As shown in FIG. 12, on the display screen of the schedule proposal output device 60, a schedule proposal 61 (priority, job name, job start time, job end time) is displayed.

  The system designer can see which job should be submitted first by looking at the schedule proposal, and can redesign the job submission procedure (schedule).

  As described above, according to the job scheduling system 1 of the present embodiment, a schedule proposal indicating the degree of influence corresponding to the subsequent job is generated and output. Therefore, a plurality of jobs are re-selected based on the degree of influence. It becomes possible to design.

  Also, by executing jobs according to the priority order that indicates the order in which the impact on subsequent jobs is large, more executable (operable) jobs can be created, so when using multiple computers, the processing time is increased. It can be shortened efficiently. Also, by creating a large number of executable jobs, without using a new high-performance computer, by using a plurality of general-purpose computers, in other words utilizing existing computing resources And processing time can be shortened.

(8) Other Embodiments In the above embodiment, although the present invention is applied to the first computer 10 and the second computer 20, the present invention is not limited to this. It can be applied to one computer (apparatus).

  In the above embodiment, the present invention is applied to a job. However, the present invention is not limited to this and can be applied to a job step in a job. In this case, the job step start time of the job step, the job step end time, and the file close information are recorded in the execution history DB 40.

  In the above embodiment, the present invention has been described as applied to a configuration (configuration A) for obtaining the impact degree evaluation value from the number of subsequent jobs that refer to a file generated by one preceding job. However, the present invention is not limited to this, and the data of the file generated by one preceding job may be configured (configuration B) to obtain the influence degree evaluation value from the number of times (number of accesses) the subsequent job group refers. According to this configuration, it is possible to cope with the case where one file is repeatedly referenced.

  In the above embodiment, although the present invention is applied to the configuration A, the present invention is not limited to this, and the scheduler can identify the configuration A and the job class (high priority for each job, etc. The definition information (C) may be combined to obtain the evaluation value of the degree of influence. In this case, the influence degree evaluation value is calculated as the influence degree evaluation value of configuration A × N.

  In the above embodiment, the present invention is applied to the configuration A. However, the present invention is not limited to this. A configuration B and a job class are combined to obtain an influence degree evaluation value (configuration D) You may In this case, the influence degree evaluation value is calculated as the influence degree evaluation value of configuration B × N.

  In the above embodiment, the present invention is applied to the configuration A. However, the present invention is not limited to this. For the configurations A to D, the subsequent job (grandchild job) of the subsequent job is further continued The second job (absorptive job) may be associated with one another to obtain an impact degree evaluation value (Configuration E).

  In the above embodiment, although the present invention is applied to the configuration A, the present invention is not limited to this, and the data amount of the subsequent job referring to the file generated by one preceding job (number x The impact evaluation value may be calculated from the number of accesses).

  In the above-described embodiment, the present invention has been described in the case where the proposed schedule is displayed on the proposed schedule output device 60. However, the present invention is not limited to this and the proposed schedule may be printed.

  In the above embodiment, although the present invention is applied to a schedule proposal including priority, job name, job start time, and job end time, the present invention is not limited to this and affects the schedule proposal. The degree evaluation value may be included, or one or more of the information may not be included.

  In the above embodiment, the present invention has been described for the case where all reference file generation flags are set to "OFF", but the present invention is not limited to this, and all the reference file generation flags are set to "OFF". Instead of the above, the file information table 80 that is set to “ON” as the verification target in the process of S31 may be the target of the count.

  1 ... job scheduling system, 10 ... first computer, 11 ... CPU, 12 ... memory, 13 ... batch program, 14 ... job schedule management section, 15 ... execution history management section, 16 ... OS, 20: second computer, 21: CPU, 22: memory, 23: schedule analysis unit, 24: schedule proposal preparation unit, 25: schedule proposal display unit, 26: OS, 30 ... ... File for batch, 40 ... Execution history DB, 50 ... Schedule proposal DB, 60 ... Schedule proposal output device.

Claims (5)

A job scheduling system used to redesign multiple job schedules, comprising:
From history information including time information indicating time when the plurality of jobs are started and access information indicating that a file is referenced or updated in each of the plurality of jobs, for each of the plurality of jobs, before and after the job Identify the relationship and specify the job that refers to the file among the jobs between the job of the starting point that has updated the file and the job of the end point that indicates the latest job that has further updated the file. A generation unit configured to set a degree of influence according to the identified job, and generate a schedule proposal in which the degree of influence is indicated;
A job scheduling system comprising: an output unit that outputs the schedule proposal. The access information includes reference number information indicating the number of times the file has been referenced;
The job scheduling system according to claim 1, wherein the generation unit counts the reference frequency information of the identified job and sets the influence degree. The generation unit is configured to determine a priority corresponding to the degree of influence of each of the plurality of jobs, a job name indicating a name of each of the plurality of jobs, and a job start time indicating a time when each of the plurality of jobs is started. Generating the proposed schedule including a job end time indicating a time when each of the plurality of jobs is ended;
The job scheduling system according to claim 1, wherein the output unit displays the priority, the job name, the job start time, and the job end time. A job scheduling method performed by a system used for redesigning a schedule of a plurality of jobs, comprising:
From history information including time information indicating time when the plurality of jobs are started and access information indicating that a file is referenced or updated in each of the plurality of jobs, for each of the plurality of jobs, before and after the job Identify the relationship and specify the job that refers to the file among the jobs between the job of the starting point that has updated the file and the job of the end point that indicates the latest job that has further updated the file. Setting a degree of influence according to the identified job, and generating a schedule proposal indicating the degree of influence;
And a second step of outputting the proposed schedule. A job scheduling apparatus used to redesign multiple job schedules, comprising:
From history information including time information indicating time when the plurality of jobs are started and access information indicating that a file is referenced or updated in each of the plurality of jobs, for each of the plurality of jobs, before and after the job Identify the relationship and specify the job that refers to the file among the jobs between the job of the starting point that has updated the file and the job of the end point that indicates the latest job that has further updated the file. A generation unit configured to set a degree of influence according to the identified job, and generate a schedule proposal in which the degree of influence is indicated;
A job scheduling apparatus comprising: an output unit that outputs the schedule proposal.

Images