JPH01188977A - Job execution control system - Google Patents

Abstract

PURPOSE:To perform due processing in its own system for the job execution control by updating or recovering the executing system information included in the job describing information for each job class based on the substitute information and therefore saving the system information on another system designated by the substitute information when the abnormality occurs in a another system. CONSTITUTION:An executing system information updating part 3 receives the n-system stop information from an another system state monitor part 2 during the start of a job. In such a case, an entry storing the job describing information 5 which shows 'n' for the executing system information 6, 'YES' for the substitute information 8, and 'wait for start' for a job status 9 based on a class-based subtable in a job start table 4 is retrieved in terms of all job classes and then a job class Ci with reception of the information on occurrence of the n-system Ci abnormality. When said information 5 is obtained, the executing system information 6 on the relevant entry is saved to the executing system saving information 7 and then changed into 'ANY'. Thus the information 6 is carried out in its own system.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention enables job execution control between multiple systems that constitute a loosely coupled computer system, and achieves effective use of system resources, improvement of throughput, and turnaround. The present invention relates to a job execution control method that makes it possible to reduce time.

(Prior Art) Conventionally, the spooling sharing function in a loosely coupled computer system has been realized as follows.

First, when a user specifies execution system information in a job control statement in a job description, this job description is held on a secondary storage device that can be shared among multiple systems. The job starting unit refers to job description information including execution system information from this secondary storage device, selects and starts a job that can be executed in its own system. Here, the execution system information is the system number of the system that executes the job, or information when load distribution is desired for effective use of system resources.

On the other hand, in a single computer system, each job is classified into job classes based on the characteristics of the system resources used by the job, and job description information for each job class is retained on a secondary storage device. It's running at an appropriate rate.

(Problem to be Solved by the Invention) The spooling function in the conventional loosely coupled computer system described above is such that jobs that are specified to be executed on another system in a job control statement are not subject to load balancing, and the other system is not running. If an abnormality occurs on another system, such as when the other system is not running, the load is heavy, or there is contention for a specific resource, the execution will be delayed for a long time until the other system's condition recovers. .

Therefore, there is a problem that the various sources of each system in the loosely coupled computer system cannot be effectively utilized.

(Means for Solving the Problems) In order to solve the above-mentioned problems, the job execution control method provided by the present invention analyzes the job description when a job is submitted, and classifies each job class based on the characteristics of the system resources to be used. a job classification function that classifies jobs into job classes and launches jobs in these job classes at an appropriate ratio, and execution system information that is information that specifies the system that executes the job as instructed by the job control statement of the job description. The job description information contained in the job description can be shared among multiple systems.
It has a spooling shared function that controls job execution among multiple systems by selecting and starting jobs that can be executed in each system based on the job description information stored in the next storage device. A method for controlling job execution in a loosely coupled computer system in which system operating information is held on a second secondary storage device that can be shared among multiple systems, the method comprising: a job starting unit that selects and starts an executable job based on the job description information classified into job classes and held and the load status of each of the job classes in its own system; Monitor the operating status of other systems based on the operating information of each of the systems held on the storage device,
The other system status monitoring unit notifies the change, and upon receiving the notification from the other system status monitoring unit, the execution system information included in the job description information held on the first secondary storage device is updated. and an execution system information update unit that updates the execution system information so that when an abnormality occurs in the own system, a job specified to be executed in the own system based on the execution system information is executed in the other system. If the execution system information is OK, change the execution system information to that effect (1), and if the other system status monitoring unit notifies that an abnormality has occurred in the other system, execute the execution on the other system for each job class. When the execution system information of the specified job is saved and then executed on the local system, and the other system status monitoring unit notifies that the other system where the error occurred has recovered. The method is characterized in that the execution system information that has been saved is recovered.

(Example) Hereinafter, an example of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a job execution control system according to an embodiment of the present invention.

The job execution control method of this embodiment includes a job starting section 1,
Other system status monitoring unit 2, execution system information updating unit 3, job startup table 4, system status table 10,
It consists of a job input section 11, an execution management section 12, and a system startup and termination section 13.

The job activation table 4 is made up of class-specific sub-tables corresponding to each job class, and the class-specific sub-tables are made up of entries that hold job description information 5 of jobs waiting to be started in each class. Job description information 5
is the execution system information 6. It consists of execution system save information 72, proxy information 8, and job status 9.

In addition, the system status table 10 holds the load status of each system as operating information, and each system has an internal table within its own system, and counts the multiplicity of its own system within this internal table. Accordingly, the load status is determined and set in the system status table 10.

This embodiment is applied to a loosely coupled computer system in which a secondary storage device is shared among multiple systems, as shown in FIG. Note that here, the own system is assumed to be the m system, and the other system is assumed to be the n system.

Next, the operation of this embodiment applied to the loosely coupled computer system shown in FIG. 2 will be explained.

The system startup and termination unit 13 sets the m-system operating flag in the system status table 10 when the system is started up, and resets it when the system is terminated.

By specifying the execution system specification control statement (SNEED statement) shown in Figure 3 in the job description, the user can specify the system number of the system on which the job is to be executed, or specify the system number on which the job is to be executed. In addition, if a system for executing the job is specified, substitute information 8 indicating whether or not it is okay to execute the job on another system when an abnormality occurs in this system is specified.

When a job is input, the job input unit 11 reads and analyzes the job description, classifies it into job classes based on the characteristics of system resources used by the job, and stores the read job description information in the job startup table 4. After registering in the class-specific sub-table of , a job input notification is sent to the job starting unit 1.

FIG. 4 is a flowchart showing the process flow of the job starting section 1.

In FIG. 4, the job starting section 1 includes a job inputting section 11.
When receiving a job input notification from (step 101), the load status of the job class to which the job belongs in its own system is obtained from an internal table. If there is sufficient load status (step 102'), and if the sub-table for each class in the job startup table 4 shows a job that can be executed by the local system, that is, the execution system information 6 is the system number "m" of the local system. or “ANY” (can be executed on any system), and if there is a registered job whose job status 9 is “Waiting to start” (
Step 103) Select and launch an appropriate job from the class-specific sub-table (Steps 104 and 105) At this time, set the job status 9 of the launched job to "Running" (Step 106) in the internal table 1 to the multiplicity counter of the relevant job class.
After adding , the load status of the own system is set in the system status table 10 (step 107). FIG. In this example, job b is started.

Further, upon receiving the job completion notification from the execution management unit 12, the job starting unit 1 deletes the job description information 5 of the job from the corresponding class sub-table in the job starting table 4 (step 108), and 1 is subtracted from the multiplicity counter of the job class, and the load status of the own system is set in the system status table (step 109).

On the other hand, as shown in FIG. 6, the other system status monitoring unit 2 periodically refers to the system status table 10 to obtain the operating status of other systems (step 201), and if the n system is not operating, (Step 202), sends an n-system stop notification to the execution system information update unit 3 (Step 203)
, when a certain job class C1 becomes overloaded (Step 206), sends an n-system C1 abnormality occurrence notification to the execution system information update unit 3 (Step 207), and
When the system starts operating (step 204), an n-system operation notification is sent to the execution system information update unit 3 (step 20).
5) When the load on class Ci becomes available, step 206) sends an n-system C1 recovery notification to the execution system information update unit 3 (step 208).

FIG. 7 is a flowchart showing the flow of execution system information update processing by the execution system information update unit 3.

The execution system information update unit 3 updates the other system status monitors 2 to n.
Upon receiving the system stop notification (Step 301), the execution system information is retrieved from the class-specific sub-table in the job startup table 4 for all job classes, and for the n-system Ci error occurrence notification (Stella 7301) for the job class C1. 6 is “n” and proxy information 8 is “Y”
ES" and the job status 9 is "Waiting for startup" (step 302). If such job description information 5 is found, the execution system information of that entry is searched. 6 to the execution system save information 7 (step 303).
, changes the execution system information 6 to "ANY" so that it can be executed on its own system (step 304).

In the example of FIG. 5, the execution system information 6 of jobs C and e is changed to become as shown in FIG. 8. A job whose execution system information 6 has been changed to "ANY" becomes a target for job activation selection by the job activation unit 1, as described earlier with reference to FIG. In the example of FIG. 8, job C and job e are selected for job activation. Here, it is assumed that job C has been selected and started.

FIG. 9 is a flowchart showing the flow of execution system information recovery processing by the execution system information update unit 3.

In FIG. 9, when the execution system information update unit 3 receives the n-system operation notification from the other-system status monitoring unit 2 (step 3
05) For all job classes, and when the n-system Ci recovery notification is received (step 305), for job class C1, the execution system save information 7 is "n" from the class-specific sub-table in the job startup table 4. Search for an entry storing job description information 5 whose job status 9 is "waiting to start" (step 306).
).

If such job description information 5 exists, the execution system information 6 is changed from the execution system retirement information 7 of the entry.
(step 307) and clear the execution system save information 7 (step 308). In the example of FIG. 8, the execution system information 6 of job e is recovered and the
The result will be as shown in the figure.

In FIG. 10, the job whose execution system information 6 has been restored to n''' becomes an n-system execution job again, and can be executed only on the n-system.

(Effects of the Invention) As explained above, the present invention updates or restores the execution system information included in the job description information for each job class based on the proxy information. One effect can be expected.

(1) If an abnormality occurs on another system, such as the other system is not running, the load is heavy, or there is contention for a specific resource, specify execution on the other system based on proxy information. By backing up the execution system information currently running and updating it to indicate that it can be executed on any system, jobs that are specified to be executed on other systems can be executed on the own system without having to wait for a long time. can be done.

(2) When the state of the other system recovers, the updated execution system information is recovered using the saved execution system information, and execution on the other system can be specified again.

(3) By flexibly executing a plurality of jobs according to the load situation of each system, the various resources of each system of the loosely coupled computer system can be utilized as effectively as possible.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing the configuration of a job execution control method according to an embodiment of the present invention, Fig. 2 is a diagram showing an example of a loosely coupled computer system, and Fig. 3 shows the syntax and parameters of an execution system specification control statement. Figure 4 is a flowchart showing the process flow of the job startup section; Figures 5, 8, and 10 are diagrams showing examples of class-based subtables in the job startup table; Figure 6; 7 is a flowchart showing the flow of the execution system information update process of the execution system information update unit. FIG. 9 is a flowchart of the execution system information recovery process of the execution system information update unit. 2 is a flowchart showing the flow. 1...Job starting section, 2...Other system status monitoring section, 3.
... Execution system information update unit, 4... Job startup table, 10... System status table, 11... Job input unit, 12... Execution management unit, 13... System startup and termination unit.

Claims (1)

[Claims] When a job is submitted, the job description is analyzed and classified into each job class according to the characteristics of the system resources to be used.
Job description information that includes a job classification function that launches jobs in these job classes at an appropriate ratio, and execution system information that is information that specifies a system that executes the job as instructed by the job control statement of the job description. is held on a first secondary storage device that can be shared among multiple systems, and jobs that can be executed by each system are selected and launched based on the job description information, thereby enabling job execution among multiple systems. A method for controlling job execution in a loosely coupled computer system that has a spooling sharing function for controlling and maintains operating information of each system on a second secondary storage device that can be shared among multiple systems. , select and start an executable job based on the job description information classified and stored in the job class on the first secondary storage device and the load status of each of the job classes in the own system; a job starting unit that monitors the operating status of the other system based on the operating information of each of the systems held on the second secondary storage device, and notifies you of changes therein; an execution system information update unit that updates the execution system information included in the job description information held on the first secondary storage device upon receiving a notification from the system status monitoring unit; If an abnormality occurs in the local system, the update unit changes the execution system information to that effect if the job specified to be executed in the local system according to the execution system information can be executed in another system. , when the other system status monitoring unit notifies that an abnormality has occurred in the other system, after saving the execution system information of the job specified to be executed on the other system for each job class; Restoring the evacuated execution system information when the execution system is changed to be executed on the own system, and when the other system status monitoring unit notifies that the other system in which the abnormality has occurred has recovered. A job execution control method characterized by: