JP2010198184A - Job management system, method for same and program for same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To execute job management allowing effective use of a hardware resource. <P>SOLUTION: This job management system includes: a job management parameter storage area 111 for storing an expected execution time and a statistical information of an execution time of a job, and storing a first table associating a load degree wherein magnitude of a load of a job execution server is digitalized, a job multiple correction degree and an execution time correction degree, and a second table associating the expected execution time and a job priority correction value; a load state decision part 112 for digitalizing the load degree based on the statistical information of the execution time and a CPU use rate of the job execution server 120, and deciding that the job execution server 120 has a high load when the load degree exceeds a predetermined threshold value; a parameter determination part 113 for determining the job multiple correction degree and the execution time correction degree in reference to the first table with the load degree as a key; and a job parameter control part 114 for changing job priority in reference to the second table with the expected execution time changed based on the execution time correction degree as a key. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a job management system, a method thereof, and a program therefor, which alleviate a high load state of a job execution server managed by the job management server and prevent occurrence of a job error due to an excess of an expected execution time caused by the high load state .

  A general job management system is required to have no influence on job execution even if a phenomenon such as a temporary increase in server load occurs. To meet this technical requirement, how to mitigate the high load state when the server is in a high load state, and how to suppress the impact of job execution delay due to the high load state There is such a problem.

  Therefore, Patent Document 1 discloses a job schedule method that dynamically changes job execution for each CPU load.

  Patent Document 2 discloses a CPU usage rate optimization method that automatically calculates a CPU usage rate to be assigned to each job in a job according to the CPU usage time of statistical information.

  Patent Document 3 discloses a load distribution method based on CPU margin management and transaction priority, in which the server determines that the server is overloaded when the CPU usage rate of the server exceeds the threshold lower limit.

JP 05-173807 A Japanese Patent Laid-Open No. 09-305417 JP 2006-1113827 A

  However, the inventions disclosed in Patent Documents 1 to 3 use hardware resources effectively only by a high CPU usage rate, so that the CPU usage rate is high or the result of excessive jobs being submitted. There is a problem that it cannot be determined whether the CPU usage rate is high.

  Even if the CPU usage rate is low, the job may be delayed due to other factors such as frequent reading or writing to the hard disk. There is also a problem that the degree of resource load cannot be accurately grasped.

  The present invention has been made in view of the above, and an object thereof is to provide a job management system capable of effectively utilizing hardware resources, a method thereof, and a program thereof.

  In order to solve the above-described problems, the job management system according to the present invention reduces the high load state of the job execution server and prevents the occurrence of a job error due to the excess of the expected execution time due to the high load state. In the job management system managed by the server, the job management server stores statistical information and estimated execution time of the job executed by the job execution server, and quantifies the load of the job execution server. The first table in which the load degree and the job multiple correction degree for correcting the job multiplicity and the execution time correction degree for correcting the expected execution time are associated with each other, and the job to be submitted to the job execution server. A job management parameter storage area for storing a second table in which job priority correction values for correcting the job priority, which is a priority of the job, are associated with each other; The load degree is quantified based on the statistical information of the execution time and the CPU usage rate of the job execution server, and when the quantified load degree exceeds a predetermined threshold, the job execution server A load state determination unit that determines that the load is high, a parameter determination unit that determines the job multiple correction level and the execution time correction level with reference to the first table using the load level as a key, and a high load For the job submitted to the job execution server determined to be, the expected execution time is changed based on the execution time correction degree, and the second table is referenced using the estimated execution time as a key. Determining a job priority correction value, and changing the job priority based on the job priority correction value; and the job multiple correction determined by the parameter determination unit. A job multiple correction degree transmitting unit that transmits the job degree to the job execution server determined to have a high load, the job execution server receiving the job multiple correction degree, and the reception And a new job multiplicity setting unit for calculating a new job multiplicity from the job multiplicity correction degree, and a job multiplicity storage area for storing the new job multiplicity.

  In order to solve the above-described problem, the job management method according to the present invention reduces the high load state of the job execution server, and prevents the job error from occurring due to the excessive execution time due to the high load state. In the job management method managed by the management server, the job management server stores statistical information on the execution time of the job executed by the job execution server and the estimated execution time, and determines the load of the job execution server. The first table in which the numerical load degree and the job multiple correction degree for correcting the job multiplicity and the execution time correction degree for correcting the expected execution time are associated with each other, and the expected execution time and the job execution server are input. A procedure for storing a second table in which job priority correction values for correcting job priorities that are job priorities are associated with each other; The load degree is quantified based on the CPU usage rate of the job execution server, and when the quantified load degree exceeds a predetermined threshold, the load for determining that the job execution server is highly loaded A status determination procedure, a parameter determination procedure for determining the job multiple correction level and the execution time correction level with reference to the first table using the load level as a key, and the job determined to have a high load For the job submitted to the execution server, the expected execution time is changed based on the execution time correction degree, and the job priority correction value is referred to by referring to the second table using the estimated execution time as a key. A job parameter control procedure for changing the job priority based on the job priority correction value, and determining that the job multiple correction degree determined in the parameter determination procedure is a high load. A job multiple correction degree transmission procedure for transmitting to the job execution server, and the job execution server receives the job multiple correction degree reception procedure for receiving the job multiple correction degree and the received job multiple correction degree. A new job multiplicity setting procedure for calculating a new job multiplicity; and a job multiplicity storage procedure for storing the new job multiplicity.

  In order to solve the above-described problem, the job management program according to the present invention reduces the high load state of the job execution server and prevents the occurrence of a job error due to excess of the expected execution time due to the high load state. In the job management program managed by the management server, the statistical information of the execution time of the job executed on the job execution server and the expected execution time are stored, and the load degree obtained by quantifying the load size of the job execution server A first table in which a job multiple correction degree for correcting a job multiplicity and an execution time correction degree for correcting the expected execution time are associated with each other, and the predicted execution time and the priority of a job input to the job execution server. A process of storing a second table in which a job priority correction value for correcting a certain job priority is associated, statistical information of the execution time, and the job execution A load state in which the load degree is quantified based on the CPU usage rate of a server and the job execution server determines that the load is high when the quantified load degree exceeds a predetermined threshold Determination processing, parameter determination processing for determining the job multiple correction degree and the execution time correction degree with reference to the first table by using the load degree as a key, and the job execution determined to be a high load For the job submitted to the server, the expected execution time is changed based on the execution time correction degree, and the job priority correction value is determined by referring to the second table using the estimated execution time as a key. A job parameter control process that determines and changes the job priority based on the job priority correction value, and the job multiple correction degree determined in the parameter determination process is determined to be a high load. A job multiple correction degree transmission process to be transmitted to the execution server is executed by the job management server, and the job multiple correction degree reception process for receiving the job multiple correction degree and a new job multiple correction degree based on the received job multiple correction degree. The job execution server is caused to execute a new job multiplicity setting process for calculating the severity and a job multiplicity storing process for storing the new job multiplicity.

  As described above, according to the present invention, a job management system capable of effectively using hardware resources by controlling the magnitude of the load of the job execution server on the basis of a numerical load level, a method thereof, and The program can be provided.

1 is a configuration diagram of a job management system according to a first embodiment of the present invention. FIG. FIG. 2 is a block diagram showing a flow of processing for setting the job multiplicity of the job execution server from the calculation of the load degree of the job management system according to the first embodiment of the present invention. FIG. 5 is a block diagram illustrating a flow of parameter setting processing when a job is submitted to a job execution server determined to be in a high load state of the job management system according to the first exemplary embodiment of the present invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[First Embodiment]
FIG. 1 is a configuration diagram of a job management system according to the first embodiment of the present invention. Referring to FIG. 1, the job management system according to the first embodiment of the present invention includes a job management server 110 and a job execution server 120.

  The job management server 110 stores statistical information on the execution time of each job and the expected execution time, and calculates the load degree obtained by quantifying the load size of the job execution server 120 and the job multiple correction degree for correcting the job multiplicity. A first table that associates an execution time correction degree that corrects an execution time, an expected execution time, and a job priority correction value that corrects a job priority that is a priority of a job input to the job execution server 120. In the job management parameter storage area 111, the job management parameter storage area 111, which is an area for storing the corresponding second table, and the high load of each job execution server 120 managed by the job management server 110 are stored in the job management parameter storage area 111. How late is the job submitted based on the statistical information of the stored execution time and the CPU usage rate of each job execution server 120 Referring to the first table stored in the job management parameter storage area 111 with the load level determined by the load state determination unit 112 as a key and the load level determined by the load status determination unit 112 as a key. A parameter determination unit 113 that determines parameters (job multiple correction degree, execution time correction degree) for correcting the degree of job multiplicity and expected execution time, and a job execution server that is determined to have a high load during degenerate operation For the job submitted to 120, the expected execution time is changed based on the execution time correction degree, and the priority is determined according to the expected execution time by referring to the second table using the estimated execution time as a key. A parameter (job priority correction value) for determining whether to change, a job parameter control unit 114 for changing the job priority based on the job priority correction value, and a parameter Comprising the job multiplicity correction of transmission means 115 for transmitting the job multiplicity correction degree determined by the over data determination unit 113 is a server determines that a high load, the.

  The job execution server 120 includes a job multiple correction degree receiving unit 121 that receives the job multiple correction degree transmitted from the job multiple correction degree transmitting unit 115, and a new job from the job multiple correction degree received by the job multiple correction degree receiving unit 121. A new job multiplicity setting unit 122 that calculates the multiplicity and stores it in the job multiplicity storage area 123 and a job multiplicity storage area 123 in which the job multiplicity of the job execution server 120 is stored.

[Operation of First Embodiment]
An operation flow of the job management system according to the first embodiment of the present invention will be described in detail with reference to the drawings.

  First, a process for setting the job multiplicity of the job execution server from the calculation of the load degree will be described with reference to FIG. FIG. 2 is a block diagram showing a flow of processing for setting the job multiplicity of the job execution server from calculation of the load degree of the job management system according to the first embodiment of the present invention.

  The load state determination unit 112 determines the load based on the CPU usage rate of each job execution server 120 managed by the job management server 110, the statistics of the execution time of the job executed on each job execution server, and the difference between the latest execution time. The information is analyzed, and a process of periodically calculating a parameter called load level of each job execution server is performed. Here, when the load level of each job execution server exceeds a predetermined threshold, the load state determination unit 112 determines that the server is in a high load state.

  The CPU usage rate of each job execution server 120 is acquired by the job management server 110 via the network as measured by an OS (Operating System) installed in each job execution server 120.

  Each job execution server 120 includes a dedicated unit for transmitting the CPU usage rate to the job management server 110. The job management server 110 may include a dedicated unit for receiving the CPU usage rate. You may use the existing communication means with which each job execution server 120 and the job management server 110 are provided as a server.

  Next, the parameter determination unit 113 refers to the first table of the load degree, job multiple correction degree, and execution time correction degree stored in the job management parameter storage area 111 for the job execution server 120 determined to be highly loaded. Then, the parameters of the job multiple correction degree and the execution time correction degree of the high-load job execution server are determined.

  The first table in which the load degree, the job multiple correction degree, and the execution time correction degree are associated with each other is determined in advance by sizing or the like and registered in the job management parameter storage area 111. An example is shown in Table 1.

次に、ジョブ多重度補正度送信手段１１５は、高負荷状態と判断されたジョブ実行サーバに、ジョブ多重補正度のパラメータを送信する。

Next, the job multiplicity correction degree transmission unit 115 transmits the job multiplicity correction degree parameter to the job execution server determined to be in the high load state.

  The job multiple correction degree receiving unit 121 receives a parameter of the job multiple correction degree.

  The new job multiplicity setting unit 122 stores the product of the current job multiplicity of the job execution server and the job multiplicity correction degree in the job multiplicity storage area 123 as the job multiplicity of the new job execution server.

  Second, a parameter setting process when a job is submitted to a job execution server determined to be in a high load state will be described with reference to FIG. FIG. 3 is a block diagram showing the flow of parameter setting processing when a job is submitted to the job execution server 120 determined to be in a high load state of the job management system according to the first embodiment of the present invention. .

  The job parameter control unit 114 sets the product of the expected execution time of the current job and the execution time correction degree as the expected execution time of the new job for the job to be input to the job execution server 120 determined to be in a high load state.

  Next, referring to the second table of parameters (job priority correction values) of how much the estimated execution time and job priority are corrected, job priority correction corresponding to the newly set expected execution time A value is acquired, and the product of the obtained job priority correction value and the priority of the input job is set as the priority of the new job. The second table in which the estimated execution time based on the load level is associated with the job priority correction value is set in advance and registered in the job management parameter storage area 111. An example is shown in Table 2.

ジョブ管理サーバ１１０は、ジョブパラメータ制御部１１４が決定したジョブ優先度に基づいて、ジョブ実行サーバ１２０にジョブを投入する。又、ジョブ実行サーバ１２０は、ジョブ多重度保存領域１２３に保存した新しいジョブ多重度に従ってジョブを処理する。これにより、ジョブ実行サーバの高負荷状態の緩和、高負荷状態が原因の予想実行時間超過によるジョブのエラー等が発生するのを防止するジョブ管理システムを実現できる。

The job management server 110 submits a job to the job execution server 120 based on the job priority determined by the job parameter control unit 114. Further, the job execution server 120 processes the job according to the new job multiplicity stored in the job multiplicity storage area 123. As a result, it is possible to realize a job management system that alleviates the high load state of the job execution server and prevents the occurrence of a job error or the like due to the expected execution time exceeding due to the high load state.

[Second Embodiment]
As a second embodiment of the present invention, when the load level is calculated by the load state determination unit 112 in FIG. 2, the load level of the job execution server 120 determined to be in a high load state is compared with the load level previously calculated. The operation is different when the load level is lower than the threshold value of the load level at which the load level is determined to be normal.

  If the load level of the job execution server determined to be in a high load state is different from the load level calculated previously, the parameter determination unit 113 again determines the load level and job multiple correction level stored in the job management parameter storage area 111. With reference to the first table of execution time correction degrees, new job multiple correction degrees and execution time correction degrees are determined.

  Thereafter, the new job multiple correction degree notified to the job execution server through the job multiple correction degree transmitting unit 115 and the job multiple correction degree receiving unit 121, and the job multiplicity originally set before the high load state is determined. Is stored in the job multiplicity storage area 123 as a new job multiplicity.

  Similarly, the job parameter control unit 114 also sets the product of the new execution time correction degree and the expected execution time of the job that was originally set before the high load state is determined as a new expected execution time. As for the job priority, a new job priority correction value is determined with reference to the second table of the estimated execution time and the job priority correction value for each new load level. Thereafter, the product of the new job priority correction value and the job priority that was originally set before the high load state is determined is set as the new job priority.

  If the load level of the job execution server 120 determined to be in the high load state is below the threshold value determined to be in the normal state, the job multiplicity of the job execution server 120 determined to be in the high load state is set to the server. The estimated execution time of the job to be submitted and the priority of the job are changed to values originally set before it is determined that the load is high. The threshold value determined as the normal state is different from the threshold value determined as the high load state, and the threshold value determined as the normal state is set lower than the threshold value determined as the high load state.

  The present invention can be implemented not only as an apparatus but also as a method and program invention.

  The present invention is suitable for job management in the reduced operation method, and can be used for a job management system capable of effectively using hardware resources in such a method.

DESCRIPTION OF SYMBOLS 110 Job management server 111 Job management parameter storage area 112 Load state determination part 113 Parameter determination part 114 Job parameter control part 115 Job multiple correction degree transmission means 120 Job execution server 121 Job multiple correction degree reception means 122 New job multiplicity setting part 123 Job multiplicity storage area

Claims (18)

In the job management system managed by the job management server so as to alleviate the high load status of the job execution server and prevent the occurrence of job errors due to excess of the expected execution time due to the high load status,
The job management server
A statistical information of an execution time of a job executed by the job execution server and an estimated execution time, and a load degree obtained by quantifying the load size of the job execution server and a job multiple correction degree for correcting the job multiplicity; A first table in which an execution time correction degree for correcting the expected execution time is associated, and a job priority correction for correcting a job priority that is a priority of a job to be input to the job execution server. A job management parameter storage area for storing a second table corresponding to values;
The load degree is quantified based on the statistical information of the execution time and the CPU usage rate of the job execution server, and when the quantified load degree exceeds a predetermined threshold, the job execution server A load state determination unit that determines that the load is high;
A parameter determination unit that determines the job multiple correction degree and the execution time correction degree with reference to the first table using the load degree as a key;
For the job submitted to the job execution server determined to have a high load, the expected execution time is changed based on the execution time correction degree and the second table is used with the estimated execution time as a key. A job parameter control unit that determines the job priority correction value and changes the job priority based on the job priority correction value;
Job multiple correction degree transmission means for transmitting the job multiple correction degree determined by the parameter determination unit to the job execution server determined to be high load;
With
The job execution server
A job multiple correction degree receiving means for receiving the job multiple correction degree;
A new job multiplicity setting unit for calculating a new job multiplicity from the received job multiplicity correction degree;
A job multiplicity storage area for storing the new job multiplicity;
A job management system comprising:   2. The load state determination unit periodically calculates the load degree from a difference between a CPU usage rate of the job execution server and statistical information of the execution time and a latest execution time. Job management system.   The job management system according to claim 1, wherein the job management server submits a job to the job execution server based on the job priority determined by the job parameter control unit.   4. The job management system according to claim 1, wherein the job execution server processes a job according to the new job multiplicity stored in the job multiplicity storage area. If the load level of the job execution server determined to be in a high load state is different from the load level calculated last time, the parameter determination unit refers to the first table to determine a new job multiple correction level and execution Determine the degree of time correction,
The job parameter control unit sets the product of the new execution time correction degree and the estimated execution time of the job that was originally set before the high load state is determined as a new estimated execution time, and the second A new job priority correction value is determined with reference to the table, and the product of the new job priority correction value and the job priority that was originally set before the high load state is determined Set as job priority,
The job execution server includes the new job multiple correction degree notified through the job multiple correction degree transmitting means and the job multiple correction degree receiving means, and the job multiplicity originally set before the high load state is determined. 5. The job management system according to claim 1, wherein the product is stored in the job multiplicity storage area as a new job multiplicity.   The job management server, when the degree of load of the job execution server determined to be a high load state is below a threshold value determined to be a normal state, the job execution server of the job execution server that has been determined to be a high load state The job multiplicity, the expected execution time of the job to be input to the job execution server, and the priority of the job are changed to the values originally set before being determined to be a high load state, and the threshold value for determining the normal state is changed. The job management system according to claim 1, wherein the job management system is set to be lower than a threshold value that is determined to be a high load state. In the job management method managed by the job management server so as to alleviate the high load state of the job execution server and prevent the occurrence of job errors due to excess of the expected execution time due to the high load state,
The job management server
A statistical information of an execution time of a job executed by the job execution server and an estimated execution time, and a load degree obtained by quantifying the load size of the job execution server and a job multiple correction degree for correcting the job multiplicity; A first table in which an execution time correction degree for correcting the expected execution time is associated, and a job priority correction for correcting a job priority that is a priority of a job to be input to the job execution server. Storing a second table with corresponding values;
The load degree is quantified based on the statistical information of the execution time and the CPU usage rate of the job execution server, and when the quantified load degree exceeds a predetermined threshold, the job execution server A load state determination procedure for determining a high load;
A parameter determination procedure for determining the job multiple correction degree and the execution time correction degree with reference to the first table using the load degree as a key;
For the job submitted to the job execution server determined to have a high load, the expected execution time is changed based on the execution time correction degree and the second table is used with the estimated execution time as a key. A job parameter control procedure for determining the job priority correction value and changing the job priority based on the job priority correction value;
A job multiple correction degree transmission procedure for transmitting the job multiple correction degree determined in the parameter determination procedure to the job execution server determined to have a high load;
With
The job execution server
A job multiple correction degree receiving procedure for receiving the job multiple correction degree;
A new job multiplicity setting procedure for calculating a new job multiplicity from the received job multiplicity correction degree;
A job multiplicity storage procedure for storing the new job multiplicity;
A method of job management, comprising:   8. The load state determination procedure according to claim 7, wherein the load degree is periodically calculated from a difference between a CPU usage rate of the job execution server and statistical information of the execution time and a latest execution time. Job management method.   The job management server according to claim 7 or 8, further comprising a procedure for submitting a job to the job execution server based on the job priority determined by the job parameter control procedure. the method of.   The job management server according to any one of claims 7 to 9, wherein the job execution server further includes a procedure of processing a job according to the new job multiplicity saved in the job multiplicity saving procedure. Method. If the load level of the job execution server determined to be in a high load state is different from the previously calculated load level, the parameter determination procedure refers to the first table to determine a new job multiple correction level and execution Determine the degree of time correction,
In the job parameter control procedure, a product of the new execution time correction degree and the expected execution time of the job that was originally set before the high load state is determined is set as a new expected execution time, and the second A new job priority correction value is determined with reference to the table, and the product of the new job priority correction value and the job priority that was originally set before the high load state is determined Set as job priority,
The job execution server includes the new job multiple correction degree notified through the job multiple correction degree transmission procedure and the job multiple correction degree reception procedure, and the job multiplicity originally set before the high load state is determined. The job management method according to claim 7, further comprising a step of storing the product of the two as a new job multiplicity.   If the load level of the job execution server determined to be in a high load state is below a threshold value determined to be in a normal state, the job multiplicity of the job execution server determined to be in a high load state, the job The estimated execution time and job priority of the job submitted to the execution server are changed to the values that were originally set before it was determined to be a high load state, and the threshold value that is determined to be a normal state is determined to be a high load state. 12. The job management method according to claim 7, wherein the job management method is set lower than a threshold value. In the job management program managed by the job management server so as to alleviate the high load state of the job execution server and prevent the occurrence of job errors due to excess of the expected execution time due to the high load state,
A statistical information of an execution time of a job executed by the job execution server and an estimated execution time, and a load degree obtained by quantifying the load size of the job execution server and a job multiple correction degree for correcting the job multiplicity; A first table in which an execution time correction degree for correcting the expected execution time is associated, and a job priority correction for correcting a job priority that is a priority of a job to be input to the job execution server. A process of storing a second table in which values are associated;
The load degree is quantified based on the statistical information of the execution time and the CPU usage rate of the job execution server, and when the quantified load degree exceeds a predetermined threshold, the job execution server A load state determination process for determining a high load;
Parameter determination processing for determining the job multiple correction degree and the execution time correction degree with reference to the first table using the load degree as a key;
For the job submitted to the job execution server determined to have a high load, the expected execution time is changed based on the execution time correction degree and the second table is used with the estimated execution time as a key. A job parameter control process for determining the job priority correction value and changing the job priority based on the job priority correction value;
A job multiple correction degree transmission process for transmitting the job multiple correction degree determined in the parameter determination process to the job execution server determined to have a high load;
To the job management server,
A job multiple correction degree reception process for receiving the job multiple correction degree;
A new job multiplicity setting process for calculating a new job multiplicity from the received job multiplicity correction degree;
A job multiplicity storage process for storing the new job multiplicity;
Is executed by the job execution server.   14. The load state determination process according to claim 13, wherein the load degree is periodically calculated from a difference between a CPU usage rate of the job execution server and statistical information of the execution time and a latest execution time. Job management program.   15. The job management according to claim 13, wherein the job management server is caused to execute a process of submitting a job to the job execution server based on the job priority determined by the job parameter control process. Program.   16. The job management according to claim 13, further comprising causing the job execution server to further execute a process of processing a job in accordance with the new job multiplicity saved in the job multiplicity saving process. Program. When the load level of the job execution server determined to be in a high load state is different from the load level calculated last time, the parameter determination process refers to the first table to determine a new job multiple correction level and execution Determine the degree of time correction,
The job parameter control process sets the product of the new execution time correction degree and the expected execution time of the job that was originally set before the high load state is determined as a new expected execution time, and the second A new job priority correction value is determined with reference to the table, and the product of the new job priority correction value and the job priority that was originally set before the high load state is determined Set as job priority,
The product of the new job multiple correction degree notified through the job multiple correction degree transmission process and the job multiple correction degree reception process and the job multiplicity originally set before being determined as a high load state is a new job. 17. The job management program according to claim 13, further causing the job execution server to execute a process of saving as multiplicity.   If the load level of the job execution server determined to be in a high load state is below a threshold value determined to be in a normal state, the job multiplicity of the job execution server determined to be in a high load state, the job The estimated execution time and job priority of the job submitted to the execution server are changed to the values that were originally set before it was determined to be a high load state, and the threshold value that is determined to be a normal state is determined to be a high load state. 18. The job management program according to claim 13, wherein the job management program is set lower than a threshold value.

Images