JP2000148512A - Method for scheduling job - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent a job from going down and to moothly use CPU resources by guaranteeing the allocation time of a CPU to respective jobs without changing the execution priority of these jobs. SOLUTION: A CPU allocation guarantee control part 12 counts CPU tirne used for a job within unit time, and when a job guaranteed at its CPU allocation does not use the CPU more than fixed time, a CPU schedular 15 preferentially allocates the CPU to the job independently of the execution priority of jobs. The preferential allocation of the CPU is appliced also to a job securing a shared resource. Even in an environment that a job having higher execution priority occupies the CPU, rninimum necessary CPU allocation is applied to respective jobs.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a job scheduling method, and more particularly, to a job scheduling method for allocating a job to a CPU of a computer system that executes a plurality of jobs in parallel.

[0002]

2. Description of the Related Art As a prior art relating to a job scheduling method for allocating a job to a CPU of a computer system which executes a plurality of jobs in parallel, for example, Japanese Patent Laid-Open No.
A technique described in Japanese Patent No. 4274 is known.

In this prior art, an execution priority is set for each job, and a CPU having a higher execution priority is given priority to a CPU.
Is assigned, or a job is submitted by designating a CPU allocation time for the job. And
In the method according to the prior art, when executing a job, only either the execution priority or the CPU allocation time can be selected.

[0004] Usually, in a computer system whose main purpose is online transaction processing, the execution priority is set in the order of an online job for executing the online transaction processing and a batch job not directly related to the online transaction processing. There is a system job that controls a computer system that has a higher execution priority than an online job. However, the system job is premised on the fact that the frequency of use of the CPU is low, so that it is considered that the CPU is preferentially assigned to the online job.

[0005]

The above-described job scheduling method in the computer system presupposes that the frequency of use of the CPU by the system job is low. However, the system job temporarily uses the CPU depending on the operating environment of the computer system. May be occupied. The prior art computer system described above comprises:
In such a case, there is a problem that even though the execution priority of the online job is set high, there is a case where the CPU is not assigned to the online job temporarily. Further, it is possible to set the online job higher than the execution priority of the system job. However, in this case, on the other hand, there is a problem that the system job cannot be operated due to the occupation of the CPU by the online job.

In the prior art described in the above-mentioned publication that specifies the CPU allocation time, the jobs to which the execution priorities are assigned and the jobs whose CPU allocation times are specified coexist. Above C
A job to which a PU is assigned has a problem that it cannot be applied to an online job because there is no request to assign a CPU to a job for which execution priority is specified. In addition, this conventional technique uses a CPU.
When a large number of jobs whose assignment time is designated are submitted,
Is occupied.

SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned problems of the prior art, and to guarantee the CPU allocation time for a job without changing the job execution priority. Even in an environment where CPUs occupy the CPU, the minimum necessary CPU is assigned to each job to prevent the job from sinking,
It is an object of the present invention to provide a job scheduling method that enables smooth operation of U resources.

[0008]

According to the present invention, there is provided a job scheduling method for allocating CPUs in accordance with job execution priorities.
This is achieved by measuring the CPU usage time of the job whose CPU allocation is to be guaranteed, and if the CPU is not allocated to the job for a predetermined time within a predetermined unit time, the CPU is preferentially allocated to the job. .

Further, the above object is achieved by the above-mentioned CPU.
When there are a large number of jobs whose allocation is guaranteed, the CPU is not continuously allocated to the jobs whose CPU allocation is guaranteed. Is satisfied, a job that is not a target of guaranteeing the allocation of the CPU holding the shared resource is set as a job whose CPU allocation is guaranteed.

According to the present invention, by providing the above-described configuration, a high execution priority is set to an online job for processing a main purpose of a computer system, and a CPU is assigned to the online job with priority over a batch job. In addition, even when a system job having a higher execution priority than an online job has a CPU assignment request, it is possible to guarantee a certain or more CPU assignment to the online job.

Further, according to the present invention, even when there are a large number of online jobs for which CPU allocation is to be guaranteed, the CPU-guaranteed online job occupies the CPU,
It is possible to prevent a situation in which a CPU is not allocated to a system job having a higher execution priority than an online job, and the online job is shared by a batch job having a lower execution priority. When waiting for resource exclusion, this batch job is also
By setting the job as a PU allocation guarantee target job, the exclusion wait of an online job can be reduced.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an embodiment of a job scheduling method according to the present invention will be described in detail with reference to the drawings.

FIG. 1 is a diagram showing the configuration of an embodiment of an apparatus for executing a job scheduling method according to the present invention, FIG. 2 is a diagram for explaining the configuration of a state holding table, and FIG.
FIG. 4 is a view for explaining the configuration of a U allocation guarantee job management table, FIG. 4 is a flowchart for explaining processing according to an embodiment of the present invention at the time of starting a job, FIG. Is a flowchart for explaining the processing of the CPU allocation assurance control unit, and FIG.
FIG. 8 is a flowchart for explaining the processing of the PU scheduler. FIG. 8 is a diagram for explaining the state of assignment of CPUs to jobs according to the embodiment of the present invention. 1 to 3, 1
1 is a timer control unit, 12 is a CPU allocation guarantee control unit,
13 is a state holding table, 14 is a shared resource exclusive control unit,
15 is a CPU scheduler, 16 is a system job, 1
7, 18 are online jobs, 19 is batch jobs, 3
3 is a job management table, and 206 is a CPU allocation guarantee job management table.

As shown in FIG. 1, the apparatus for executing a job schedule according to an embodiment of the present invention includes a timer control unit 11, a CPU allocation guarantee control unit 12, a state holding table 13, a shared resource exclusive control, Unit 14 and CPU
The online jobs A17 and B1 which are configured by the scheduler 15 and are previously guaranteed to be assigned CPUs.
8 and the other system jobs 16 and batch jobs 19 are subjected to CPU assignment processing.

In the above, the timer control unit 11 activates the CPU allocation guarantee control unit 12 at regular time intervals, and
The U allocation assurance control unit 12 includes an online job A17 and an online job B of the CPU allocation guaranteed job.
The CPU allocation time is calculated with the CPU 18, and if a CPU longer than a predetermined time has not been allocated within a predetermined period, a priority schedule request for the job to be allocated to the CPU is issued to the state holding table 13.

When there is a priority schedule request in the status holding table 13, the CPU scheduler 15 performs a process of allocating a CPU to the job more preferentially than the system job 16 having a higher execution priority. Further, the shared resource exclusion control unit 14 controls the batch job 1 having a low execution priority order.
When the CPU allocation guaranteed job requests the shared resource used by the job 9, the batch job 19 using the shared resource is also processed to be the CPU allocation guaranteed job.

As shown in FIG. 2, the state holding table 13 is provided with an area for storing various information described below. The timer activation interval 202 defines a time interval at which the timer control unit 11 activates the CPU allocation guarantee control unit 12, for example, 15 ms. The CPU allocation unit monitoring time 203 is a CPU allocation monitoring time for determining whether or not a job is to be subject to CPU allocation guarantee, for example, 1 s.
When the CPU is not assigned to the job for the time set in the CPU assignment guarantee time 204 within this time interval, for example, 100 ms, it is used to preferentially assign the CPU to the job. Is done.
The CPU assignment time 205 defines a priority assignment time when assigning a CPU preferentially, for example, 10 ms. CPU allocation guarantee job management table 206
Is a table for managing a job for which CPU allocation is guaranteed, and its details will be described later with reference to FIG.

Job management area 2 for which CPU allocation is guaranteed
07 defines a job for which CPU allocation is guaranteed. The priority schedule queue pointer 208 is
This is an area for queuing a job management table of a job to which a CPU is to be preferentially allocated to guarantee CPU allocation. The priority schedule request area 209 stores C
When preferentially assigning a CPU to a PU allocation guaranteed job, the priority job is assigned to the CPU scheduler 1
5 is a table for contacting No. 5. The in-priority schedule display flag 210 is a flag for the CPU allocation assurance control unit 12 to identify whether or not the CPU is being preferentially allocated to the job.

CPU assignment guarantee job management table 2
Reference numeral 06, as shown in FIG. 3, includes a job control table head pointer 32 and a job management table 33 created for each CPU allocation guaranteed job. The job control table head pointer 32 is a job management table 33 provided for each CPU allocation guaranteed job.
Is a pointer for storing the start address of the table at the head of. The job management table 33 includes an area for storing various information described below.

The NEXT pointer 34 is a pointer to the next CPU allocation guarantee management table 33. The job ID 35 is an identification number for identifying a job. The monitoring mode flag 36 is a flag indicating that the CPU allocation time per unit time is being measured when the CPU allocation guarantee target job is in the CPU allocation waiting state, and the monitoring start time 37 is the time when the monitoring mode is set. Is an area for storing. The CPU usage time saving area 38 at the start of monitoring is an area for saving the CPU usage time used by a job up to that time when the monitoring mode is set to obtain the CPU allocation time per unit time. The temporary priority flag 39 is a flag indicating that the job holding the shared resource has been temporarily set as the CPU allocation guaranteed job when the CPU allocation guaranteed job enters the exclusion wait state for the resource shared.

Next, the processing at the start of a job will be described with reference to the flow shown in FIG.

(1) The CPU allocation guarantee control unit 12
At the start of a job, that is, when a job is input to the computer system, it is determined whether or not the job is a CPU allocation guaranteed job by referring to the CPU allocation guaranteed job management area 207 of the status holding table 13 ( Step 41).

(2) If it is determined in step 41 that the input job is the batch job 19 or the like and is not a CPU allocation assurance target job, the CPU allocation assurance control unit 12 performs processing without doing anything. finish.

(3) If it is determined in step 41 that the input job is the online job A 107 or B 108 or the like and is a job targeted for CPU allocation guarantee, the CPU allocation guarantee control unit 12 A job management table 33 for guaranteeing the job is created (step 42).

(4) The CPU allocation guarantee control unit 12
The job management table 33 created in step 42 is
It is registered in the PU allocation guarantee job management table 206 (step 43).

Next, the processing of the shared resource exclusive control unit 14 will be described with reference to the flow shown in FIG.

(1) In the shared resource securing request processing, it is determined whether or not the submitted job is waiting for shared resources.
If it does not wait for the shared resource, no processing is performed (step 51).

(2) If it is determined in step 51 that the input job is waiting for a shared resource, it is determined whether or not the resource requesting job is a CPU allocation guaranteed job.
If the resource requesting job is not a CPU allocation guarantee target job, no processing is performed (step 52).

(3) If it is determined in step 52 that the resource request source job is the online job A17 or the online job B18 which is the job for which CPU allocation is guaranteed, the job securing the shared resource is the job for which CPU allocation is guaranteed. It is determined whether the job is a job or not, and if the job securing the shared resource is a CPU allocation guaranteed job, no processing is performed (step 53).

(4) In step 53, if the job securing the shared resource is the batch job 19 that is not a CPU allocation guaranteed job, the batch job securing the shared resource is temporarily set as the CPU allocation targeted job. To create the job management table 33 (step 54) (5) The job management table 33 created in step 54
Is set to the temporary priority flag 39, and the job management table 33
(Steps 55 and 56).

With the above processing, a batch job for which a shared resource has been secured can be temporarily set as a CPU allocation guaranteed job.

(6) In the shared resource release request processing, when the job releases the shared resource, the job
It is determined whether the job is a PU allocation guarantee job by checking the value of the temporary priority flag 39 of the job management table 33, and if the temporary priority flag 39 is not set, no processing is performed (step 57).

(7) If it is determined in step 57 that the batch job is a temporary CPU allocation guaranteed batch job, the job management table 3 of the batch job whose shared resources have been released
3 is deleted from the CPU assignment guarantee job management table 206, and the job management table 33 is deleted (steps 58 and 59).

By the above processing, the status of the batch job for which the temporary CPU allocation is guaranteed can be released.

Next, referring to the flow shown in FIG.
The processing of the U allocation guarantee control unit 12 will be described.

(1) When started by the timer control unit 11, first, the CPU allocation guarantee job management table 20
The first job management table is obtained from the job control table top pointer 32 of No. 6 and it is determined whether or not the job management table 33 exists (steps 601 and 602).

(2) If it is determined in step 602 that there is a job management table 33, the job
Whether or not to wait for assignment is determined based on information held by the OS,
When the job is waiting for CPU assignment, it is determined whether the job is in the CPU assignment time monitoring mode. Note that the aforementioned OS
May be stored in the job management table 33 as a flag indicating whether or not it is waiting for CPU allocation (steps 603 and 604).

(3) If it is determined in step 604 that the operation mode is the CPU allocation monitoring mode, the job has a CPU allocation guarantee time within the unit time from the start of monitoring.
It is determined whether the PU has been used. CPU usage time is
The job management table 33 is calculated based on the total CPU usage time from the start of the job held in the OS.
Is obtained as a value obtained by subtracting the time stored in the CPU use time save area 38 at the time of the start of monitoring. In the above-described determination, when the CPU for the CPU allocation guarantee time is not used, it is determined whether the CPU allocation unit monitoring time has elapsed from the monitoring start time (steps 605 and 60).
6).

(4) If it is determined in step 606 that the CPU allocation unit monitoring time has elapsed, a job management table is registered in a priority schedule queue in order to preferentially allocate a CPU to the job. By obtaining a management table and repeating the processing of step 602 and subsequent steps, processing for all CPU allocation guaranteed jobs is performed (steps 607 and 608).

(5) If it is determined in step 604 that the current mode is not the CPU allocation time monitoring mode, and if it is determined in step 605 that a CPU longer than the guaranteed CPU allocation time has been used per unit time, the CPU allocation time Is set to monitor the CPU allocation time monitoring mode flag (step 610).

(6) Thereafter, the CPU usage time is saved,
After saving the start time of the CPU allocation time monitoring, the processing shifts to the processing from step 608 described above (step 6).
11, 612).

(7) If it is determined in step 603 that the job is not in the CPU allocation waiting state because the job is being executed or there is no processing to be performed, the CPU allocation time monitoring is performed to release the monitoring state. The mode flag is reset, and the process proceeds to the processing from step 608 described above (step 609).

(8) When it is determined in step 602 that there is no job management table, that is, after all job management tables have been processed, it is determined whether or not the job is in the priority schedule (step 613).

(9) If it is determined in step 613 that the priority scheduling of the job is not being performed, it is determined whether or not the job management table of the job to which the CPU is to be assigned priority is registered in the priority scheduling queue (step 613). 614).

(10) If it is determined in step 614 that the job management table of the job to which the CPU is to be preferentially assigned is registered in the priority schedule queue, the priority schedule display area is set, and the priority schedule request area is set. By setting a job ID to the job scheduler, the CPU scheduler 15 is requested to give a job priority schedule (steps 615 and 616).

(11) After requesting the priority schedule in step 616, the job management table of the job to be priority-scheduled is deleted from the priority schedule queue, thereby preventing the same job from being successively scheduled. (Step 617).

(12) If it is determined in step 613 that the priority scheduling of the job is being performed, the priority schedule request area is reset. Thus, the CPU scheduler 15 allocates the CPUs according to the execution priority, and assigns the CPU priority. This prevents the target job from occupying the CPU. If it is determined in step 614 that the job management table of the job to which the CPU should be preferentially assigned to the priority schedule queue is not registered, the priority schedule display is canceled and the priority schedule request area is reset. As a result, the CPU scheduler 15 can execute the CPU allocation according to the execution priority (steps 618, 619).

Next, referring to the flow shown in FIG.
The process of the U scheduler 15 will be described.

(1) First, it is determined whether or not the job ID of the job to be scheduled for priority is set in the priority schedule request area 209 of the status holding table 13 (step 71).

(2) If it is determined in step 71 that a job ID has been set in the priority schedule request area 209, it is determined whether or not this job is waiting for CPU allocation. The CPU is assigned to the job (steps 72 and 73).

(3) In step 72, if the priority schedule request job is not in the CPU assignment waiting state, the priority schedule request area is reset, a CPU assignment waiting job is selected according to the job execution priority, and the CPU is assigned to the job. Assignment (steps 74 and 75).

(4) If it is determined in step 71 that the job ID is not set in the priority schedule request area 209, a CPU assignment waiting job is selected in accordance with the job execution priority, and the CPU is assigned to the job.

Next, with reference to FIG. 8, a description will be given of the status of the assignment of the CPU to the job according to the embodiment of the present invention described above.

FIG. 8A is a diagram for explaining the concept of CPU allocation when the present invention is not applied. When a situation such as a system job having a high execution priority enters a loop state occurs, Occupies the CPU. In this case, as a matter of course, the online job A and the online job B have C
No PU will be assigned.

FIG. 8B is a diagram for explaining the concept of CPU allocation to each job according to the embodiment of the present invention. As can be seen from this figure, the online job A and the online job B, which are CPU allocation guaranteed jobs,
The CPU is also allocated when a system job having a higher execution priority enters a loop state. Then, no CPU is assigned to a batch job whose execution priority is low and is not a CPU assignment guarantee target. Further, in this example, since the CPU allocation guaranteed job does not occupy the CPU, the system job can also operate.

FIG. 8C is a diagram for explaining the concept of CPU allocation when an exclusion wait of a shared resource occurs according to the embodiment of the present invention. The illustrated example shows a case in which the online job A of the job whose CPU allocation is to be guaranteed has waited for the shared resource secured by the batch job.
In this case, the batch job is temporarily set as the CPU allocation guarantee target job, so that even if the system job or online job B enters a loop state,
PUs are allocated and batch jobs can execute processing. As a result, since the batch job releases the shared resource, the CPU can be assigned to the online job A waiting for the shared resource at this time. At this time, the CPU assignment guarantee of the batch job temporarily set as the CPU assignment guarantee target job is released, so that the CPU is not preferentially assigned to the batch job.

It should be noted that the CPU priority assignment time shown in FIG.
Is the time set as Further, in FIG. 8C, the reason that the online job A has finished processing without using up the CPU allocation time is that the shared job waits for the shared resource at this time, and the remaining CPU allocation time is The job to which the next CPU allocation time is allocated, in the illustrated example, is allocated to the system job. Also, the remaining CPU allocation time after the batch job releases the shared resources is allocated to the system job in the same manner as described above.

According to the embodiment of the present invention described above, the CP
Since the CPU can be preferentially assigned without changing the execution priority of the U allocation guaranteed job, the CPU can be scheduled according to the execution priority when the CPU does not need to be assigned priority.
Further, according to the above-described embodiment of the present invention, the CPU scheduler does not allocate a CPU only to a job for which CPU allocation is guaranteed, so that the CPU is not occupied by a job for which CPU allocation is guaranteed. . Further, according to the embodiment of the present invention described above, the CP
When the U allocation guaranteed job is waiting for shared resources, the job securing the shared resources is temporarily set as the CPU allocation guaranteed job, so that the sinking of the CPU allocation guaranteed job due to waiting for the shared resources is prevented. can do.

In the above-described embodiment of the present invention, the configuration and the control method of the status holding table 13 and the CPU assignment assurance job management table 206 can be variously modified without departing from the gist of the present invention. it can.
Further, the various time values exemplified in the description of the embodiment of the present invention are not limited to the values and can be set arbitrarily.

[0060]

As described above, according to the present invention, the CPU time allocated to a job within a certain unit time is guaranteed without changing the execution priority of the job. When a proper CPU is allocated, the CPU can be allocated according to the execution priority. Further, according to the present invention, even when there are many jobs for which the CPU allocation time is to be guaranteed, it is possible to prevent these jobs from occupying the CPU. Further, according to the present invention, when a job whose CPU allocation time is to be guaranteed is waiting for the exclusion of a shared resource, the CPU allocation is guaranteed for the job which has secured the shared resource. The exclusion wait can be reduced, the CPU allocation to the job can be guaranteed, and the CPU resources of the computer system can be operated smoothly.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a configuration of an embodiment of an apparatus that executes a job scheduling method according to the present invention.

FIG. 2 is a diagram illustrating a configuration of a state holding table.

FIG. 3 is a diagram illustrating a configuration of a CPU assignment guarantee job management table.

FIG. 4 is a flowchart illustrating a process according to an embodiment of the present invention when a job is started.

FIG. 5 is a flowchart illustrating processing of a shared resource exclusion control unit.

FIG. 6 is a flowchart illustrating a process of a CPU allocation guarantee control unit.

FIG. 7 is a flowchart illustrating processing of a CPU scheduler.

FIG. 8 is a diagram illustrating a state of assignment of a CPU to a job according to the embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 Timer control part 12 CPU allocation guarantee control part 13 State holding table 14 Shared resource exclusion control part 15 CPU scheduler 16 System job 17, 18 Online job A, B 19 Batch job 206 CPU allocation guarantee job management table

 ──────────────────────────────────────────────────の Continuing on the front page (72) Inventor Chieko Akiba 5030 Totsuka-cho, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture Inside the Software Division of Hitachi, Ltd. (72) Inventor Ichiro Ishioka 5030 Totsuka-cho, Totsuka-ku, Yokohama-shi, Kanagawa Hitachi, Ltd. Software Division F-term (reference) 5B098 CC03 GA03 GB01 GB13 GC01 GC03 GD02 GD14

Claims (3)

[Claims] 1. A job scheduling method for allocating a CPU according to a job execution priority.
A job to be guaranteed for CPU allocation that guarantees the assignment of
A method of scheduling a job, comprising measuring usage time, and assigning a CPU to the job preferentially if the job has not been assigned a predetermined time or more within a predetermined unit time. 2. The job scheduling method according to claim 1, wherein when there are a large number of CPU allocation guaranteed jobs, CPUs are not continuously allocated to the CPU allocation guaranteed jobs. 3. When a job whose CPU allocation is to be guaranteed is waiting for exclusion of a resource shared between jobs, a job that is not guaranteed to be allocated by a CPU holding a shared resource is replaced by a CPU allocation guaranteed job. 3. The job scheduling method according to claim 1, wherein: