JPH07175766A - Job reexecution control method for loosely-coupled multiplexing system - Google Patents

Abstract

PURPOSE:To promptly perform the reexecution of a job in execution by a designated host computer Hm when a fault is generated in a host computer Hn. CONSTITUTION:In a host computer H1, a job control language translation means B2 translates a job control language B1 and registers the designation of a host computer Hm reexecuting the job of a host computer Hn in a job control information holding means A4. A host fault recognition means A1 recognizes the fault notification from a host monitoring device G1, and a job reexecution preparation means A2 updates the job control information of the job control information holding means A4 and requests the reexecution for the job in execution in a faulty host computer Hn. A job schedule means A3 performs the rescheduling of the job and requests the host computer Hm to reexecute the job. A job start means A5 starts the execution program A6 of the job for which the execution is requested.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a job re-execution control method for a loosely coupled multiple system, and more particularly to a job re-execution control method for a loosely coupled multiple system when a failure occurs in a host computer that is executing a job.

[0002]

2. Description of the Related Art A conventional job re-execution control method for a loosely coupled multiple system is to re-execute a job being executed when a failure occurs in the host computer that is executing the job. By holding the necessary job control information in the secondary storage area of the host computer, after the failure of the failed host computer is recovered, the contents of the job control information in the secondary storage area are used to interrupt due to the failure. The executed job is automatically re-executed.

As an example of such a conventional job re-execution control system for a loosely coupled multiplex system, Japanese Patent Application Laid-Open No. 62-795.
31. "Job step restart method".

[0004]

The above-mentioned conventional job re-execution control method for the loosely coupled multiplex system, when a failure occurs in the host computer executing the job,
After the failure of the host computer is recovered, the job interrupted by the failure is re-executed due to the contents of the job control information in the secondary storage area of the host computer, so until the failure of the host computer is recovered However, there is a drawback that the re-execution of the job whose execution has been interrupted cannot be executed, and the operator must give an instruction for re-execution of the job after the failure of the host computer is recovered.

Further, according to the conventional job re-execution control method of the loosely coupled multiple system, there is a case where it is difficult to recover the failure of the host computer due to the failure of the host computer in which the failure has occurred and the job cannot be re-executed. It has the drawback of being.

An object of the present invention is to, when a failure occurs in a host computer executing a job, promptly re-execute the job by another host computer without waiting for recovery from the failure of the host computer. It is to provide a job re-execution control method of a loosely coupled multiple system that can be performed.

[0007]

According to the job re-execution control method of the loosely coupled multiplex system of the first invention, a failure occurs in any one of the host computers by connecting to each of the plurality of host computers. In the job re-execution control method of the loosely coupled multi-system having a host monitoring device that notifies other host computers that a failure has occurred, the first host computer includes (A) the host computer Job management information holding means for holding job control information for a job entered in (B) a job control language is input and translated, and when a failure occurs in the second host computer due to the job control language If a third host computer to re-execute the job is specified, the Job control information translating means for registering the job control information as job control information in the job management information holding means, and (C) when a failure occurs, the host monitoring apparatus is notified of and recognizes the second host computer in which the failure occurred By referring to the job control information of the job management information holding means, a second host failure recognition means for making a re-execution request for a job being executed in the host computer; and (D) a job from the host failure recognition means Job re-execution preparation means for making a job re-scheduling request to the third host computer designated by the job control information by referring to the job control information of the job management information holding means by receiving the job re-execution request. (E) by receiving the job control information from the job control language translation means, Job to be executed by the host computer, and the job is rescheduled in response to a job rescheduling request from the job reexecution preparation means to the third host computer. And a second job scheduling means for requesting re-execution of the job of the host computer, and each of the plurality of host computers includes (F) a job based on the schedule from the job scheduling means. Job execution means for activating an execution program of a job requested to be executed when there is an execution request is configured.

The job re-execution control method for the loosely coupled multiple system according to the second aspect of the present invention is the job re-execution control method for the loosely coupled multiple system according to the first aspect of the present invention. The re-execution of the job when the host computer according to claim 1 for re-execution is not specified or the specified specification is changed and a failure occurs in the second host computer. And a terminal for inputting the designation of the third host computer to perform the job, and (B) a third host computer to re-execute the job when a failure occurs in the second host computer from the terminal. If you do
Job re-execution host registration means for registering the designation as new job control information in the job management information holding means is provided for the first host computer.

A job re-execution control method for a loosely coupled multiple system according to a third aspect of the present invention is the job re-execution control method for a loosely coupled multiple system according to the first aspect of the present invention. A user management information storage means for storing the designation of the host computer according to claim 1 for re-execution for each user, and (B) the user management information storage means for the user of the faulty job. 2. The job management information according to claim 1, wherein when a failure occurs in the user management information acquisition means for acquiring the designation of the host computer for re-executing the job from (C) the second host computer. If the host computer for re-execution of the job is not designated in the holding means, the user of the job in failure is given to the user management information acquisition means. 2. The host according to claim 1, wherein the designation of the host computer for the user is acquired, the designation is registered as new job control information in the job management information holding unit, and a re-execution request is made for the faulty job. A fault recognition means is provided for the first host computer.

A job re-execution control method for a loosely coupled multiple system according to a fourth aspect of the present invention is the job re-execution control method for a loosely coupled multiple system according to the first aspect of the present invention. A system definition information holding unit for holding the designation of the host computer according to claim 1 for re-execution, (B)
Host information acquisition means for acquiring a designation of the host computer for re-executing a job from the system definition information holding means for the second host computer in which a failure has occurred; and (C) the second host It is the second host computer when the host computer for re-execution of the job is not designated in the job management information holding means according to claim 1 when the computer fails. To the host information acquisition means, the designation of the third host computer with respect to the second host computer is obtained, and the designation is registered in the job management information holding means as new job control information. A host failure recognizing means according to claim 1, wherein a re-execution request is made to the failed job together with the first job. And it is configured to include to the host computer.

Further, the job re-execution control method for the loosely coupled multiple system according to the fifth aspect of the present invention is the job re-execution control method for the loosely coupled multiple system according to the first aspect of the present invention. The system definition information holding means for holding the designation of the host computer according to claim 1 for re-execution for each job class, and (B) the system definition information holding means for the job class of the faulty job. 2. The job management information holding according to claim 1, wherein the job class information acquisition means for acquiring the designation of the host computer for re-execution of the job of 1) and (C) the second host computer when a failure occurs. If the host computer is not specified to re-execute the job, the job class for the job in error By giving to the job class information acquisition means, the designation of the third host computer for the job class is obtained, the designation is registered as new job control information in the job management information holding means, and The host failure recognizing means according to claim 1, which issues a re-execution request for a job, is provided for the first host computer.

[0012]

Embodiments of the present invention will now be described with reference to the drawings. FIG. 1 is a block diagram showing a first embodiment of a job re-execution control method for a loosely coupled multiplex system according to the present invention. In the first embodiment, as shown in FIG. 1, a plurality of host computers H1, ... Hm, ... Hn, ... And a host monitoring device connected to each of the host computers H1 ,. And G1.

In the following description, the host computer H1 will be described as a job scheduling host, the host computer Hn as a failed host, and the host computer Hm as a job re-execution host. And

Further, the job control language B1 specifies the host computer Hm which should re-execute a job when a failure occurs in the host computer Hn, and is preloaded in the host computer H1.

On the other hand, the host computer H1 has a host failure recognition means A1 and a job re-execution preparation means A2.
A job scheduling means A3, a job management information holding means A4, and a job control language translation means B2.

Further, the host computers H1, ... H
Each of m, ... Hn, ... Has a job starting means A5. Further, the job activation means A5 activates the execution program A6 of the designated job according to the instruction from the job scheduling means A3.

FIG. 2 is a flow chart showing an example of the operation at the time of job input processing in the first embodiment. Also, FIG.
6 is a flow chart showing an example of an operation when a failure occurs in the host computer in the first embodiment.

When a job is submitted in the first embodiment, the job control language translation means B2 sequentially inputs and translates the job control language B1 and the like to translate the job submitter, the job execution class, and the job execution class. Host computer H
1, ... Hm, ... Hn, ... are determined and registered in the job management information holding means A4, and then the job scheduling means A
3 has been notified.

Therefore, the job scheduling means A3
Upon receiving the notification from the job control language translation means B2, the job management information holding means A4 is searched and the host computer H
, ... Hm, ... Hn, ... Schedule jobs to be executed, and if there are executable jobs,
Host computers H1, ... H that execute the job
An activation request is issued to any of m, ... Hn ,. Upon receiving the activation request, the job activation unit A5 activates the execution program A6 of the designated job according to the instruction from the job scheduling unit A3.

Then, as shown in FIG.
If the host computer Hm that should re-execute the job being executed is designated by the job control language B1 in step 1 when the host computer Hn fails, the designation is made in step 202. It is registered in the job management information holding means A4 as job control information. Further, in step 203, the job scheduling means A3 receives the job control information from the job control language translating means B2, so that the host computers H1, ...
Each job to be executed is scheduled in each of Hm, ... Hn ,.

On the other hand, in step 301, the host failure recognizing means A1 is notified by the host monitoring device G1 that a failure has occurred in the host computer Hn when the host computer Hn has failed. The job control information registered in the job management information holding unit A4 is checked, and when the job is being executed by the host computer Hn in which the failure has occurred, in step 303, the re-execution host computer Hm is designated. When it is detected, the job re-execution preparation unit A2 is requested to perform a process for re-execution of the job, and the process proceeds to step 304.

Then, the job re-execution preparation means A2 prepares for re-execution of the job in step 304.
The resource reserved for the faulty job is released, the contents of the job management information holding unit A4 is returned to the state before the execution of the job is executed in step 305, and the execution of the job is executed in step 306. To the host computer Hm for re-execution, and the job scheduling means A3
Request to re-execute the job.

Therefore, the job scheduling means A3
In step 307, the job re-execution preparation unit A2 receives a job re-scheduling request for the host computer Hm, and re-scheduling is performed based on the contents of the job management information holding unit A4. Since the computer Hn requests the re-execution of the job being executed, the job activating means A5 of the host computer Hm activates the execution program A6 of the job for which the re-execution request is received.

FIG. 4 is a block diagram showing a second embodiment of the job re-execution control system of the loosely coupled multiplex system according to the present invention. In the second embodiment, as shown in FIG. 4, a plurality of host computers H11, ... H1m, ... H1n, ... And host computers H11, ... H1m ,.
Connected to host computer H11, ... H1m, ... H
If any one of 1n, ... Has a failure, the host monitoring device G11 is provided for notifying that the failure has occurred.

Further, the job control language translation means B12 of the host computer H11 inputs and translates the job control language B11 and the like and registers it in the job management information holding means A14. ,
Host computers H11, ... H1m, ... H1n, ...
It holds job control information for the jobs submitted in it. The job control language translation unit B12 uses the job control language B11 as job control information when the host computer H1m that should re-execute the job is designated when a failure occurs in the host computer H1n. It is registered in the management information holding means A14.

On the other hand, in the terminal C11 of the host computer H11, when the job management information holding means A14 fails, the host computer to which the job should be re-executed is not specified or is executed. When the user wants to change the designation, when the host computer H1n fails, the designation of the host computer H1m to re-execute the job is input.

The terminal C of the host computer H11
The host computer H1 that should re-execute the job when a failure occurs from 11 to the host computer H1n
When m is specified, the job re-execution host registration means C1
2 has registered the designation as job control information in the job management information holding unit A14.

Therefore, when a failure occurs, the host failure recognition means A11 of the host computer H11 causes the host monitoring apparatus G11 to cause a failure in the host computer H1.
By notifying and recognizing n, the job control information of the job management information holding unit A14 is referred to, and a request for re-execution of the job being executed by the host computer H1n is made.

The job re-execution preparation means A12 of the host computer H11 is then connected to the host failure recognition means A11.
In response to the request for re-execution of the job from the job control information holding unit A14, the job re-scheduling request is issued to the host computer H1m designated by the job control information.

On the other hand, the job scheduling means A13 of the host computer H11 is the job control language translation means B.
.. H1m, ... H1n, ... Scheduling jobs to be executed by the host computers H11, ... H1m ,. When the job is received, the job is rescheduled to request the host computer H1m to reexecute the job of the host computer H1n.

Therefore, when there is a job execution request based on the schedule from the job scheduling means A13, the job starting means A15 of the host computer H1m.
Has started the execution program A16 of the job requested to be executed.

FIG. 5 is a block diagram showing a third embodiment of the job re-execution control system of the loosely coupled multiplex system according to the present invention. In the third embodiment, as shown in FIG. 5, a plurality of host computers H21, ... H2m, ... H2n, ... And host computers H21, ... H2m ,.
To host computer H21, ... H2m, ... H
When any one of 2n, ... Has a fault, the host monitoring device G21 notifies the fact that the fault has occurred.

Since the job control language translation means B22 of the host computer H21 inputs and translates the job control language B21 and the like and registers it in the job management information holding means A24, the job management information holding means A24 is ,
Host computer H21, ... H2m, ... H2n, ...
It holds job control information for the jobs submitted in it. Then, when the host computer H2n that should re-execute the job when the host computer H2n fails due to the job control language B21, the job control language translation means B22 uses the designation as job control information. It is registered in the management information holding means A24.

On the other hand, the user management information storage means D21 of the host computer H21 includes host computers H21, ... H2m, ... H2n, ... For re-executing a job when a failure occurs for each user. The user management information acquisition unit D22, which stores the designation, re-executes the job from the user management information storage unit D21 for the user of the faulty job.
The designations of H2m, ... H2n, ... Can be acquired.

Therefore, when a failure occurs, the host failure recognition means A21 of the host computer H21 causes the host monitoring apparatus G21 to cause a failure in the host computer H2.
By being notified and recognizing n, the host computer H21, ... H2m, ... H2n, for re-executing the job in the job management information holding unit A24 by referring to the job control information of the job management information holding unit A24, .. is not specified, the user of the faulty job is given to the user management information acquisition means D22 to specify the host computer H2m for re-executing the job for that user. The job management information holding unit A2 that acquires the designation and uses the designation as new job control information
No. 4 is registered and the re-execution request is made for the faulty job being executed by the host computer H2n.

Then, the job re-execution preparation means A22 of the host computer H21 uses the host failure recognition means A21.
In response to the request for re-execution of the job from the job control information holding unit A24, the job re-scheduling request is issued to the host computer H2m designated by the job control information.

On the other hand, the job scheduling means A23 of the host computer H21 has a job control language translation means B.
.. H2m, ... H2n, .. Schedule the job to be executed by the host computers H21, ... H2m ,. When the job is received, the job is rescheduled to request the host computer H2m to re-execute the job of the host computer H2n.

Therefore, when there is a job execution request based on the schedule from the job scheduling means A23, the job starting means A25 of the host computer H2m.
Has started the execution program A26 of the job requested to be executed.

FIG. 6 is a block diagram showing a fourth embodiment of the job re-execution control system of the loosely coupled multiplex system according to the present invention. In the fourth embodiment, as shown in FIG. 6, a plurality of host computers H31, ... H3m, ... H3n, ... And host computers H31, ... H3m ,.
Connected to host computer H31, ... H3m, ... H
If any of the 3n, ... Has a fault, the host monitoring device G31 for notifying that the fault has occurred is configured.

Since the job control language translation means B32 of the host computer H31 inputs and translates the job control language B31 and the like and registers it in the job management information holding means A34, the job management information holding means A34 ,
Host computer H31, ... H3m, ... H3n, ...
It holds job control information for the jobs submitted in it. Then, when the host computer H3n that should re-execute the job when the host computer H3n fails due to the job control language B31, the job control language translation unit B32 uses the designation as the job control information for the job. It is registered in the management information holding means A34.

On the other hand, the system definition information holding means F31 of the host computer H31 holds the designation of the host computer for re-executing the job when a failure occurs for each host computer and for each job class. Therefore, the host information acquisition unit E31 acquires the designation of the host computer H3m that should re-execute the job from the system definition information holding unit F31 for the failed host computer H3n, and the job class information acquisition unit E32 The designation of the host computer H3m for re-executing the job is acquired from the system definition information holding unit F31 for the job class of the job in which the failure has occurred.

Therefore, when a failure occurs, the host failure recognition means A31 of the host computer H31 causes the host monitoring apparatus G31 to cause a failure in the host computer H3.
When n is notified and recognized to refer to the job control information of the job management information holding unit A34, and the host computer for re-executing the job is not specified in the job management information holding unit A34 In addition, by giving the host information acquisition means E31 that the failure is the host computer H3n, the designation of the host computer H3m corresponding to the host computer H3n is acquired, or the job class of the failed job is acquired as the job class information. The designation is given to the means E32 to obtain the designation of the host computer H3m for the job class, and the designation is registered in the job management information holding means A34 as new job control information and executed by the host computer H3n. To the job of the fault that was It is doing the re-execution request.

The job re-execution preparation means A32 of the host computer H31 is then connected to the host failure recognition means A31.
In response to the job re-execution request from the job control information holding unit A34, the job re-scheduling request is issued to the host computer H3m designated by the job control information.

On the other hand, the job scheduling means A33 of the host computer H31 is the job control language translation means B.
.. H3m, ... H3n, ... Scheduling of the job to be executed by the host computers H31, ... H3m, ..., And the job re-scheduling request from the job re-execution preparation means A32 to the host computer H3m. When the job is received, the job is rescheduled to request the host computer H3m to reexecute the job of the host computer H3n.

Therefore, when there is a job execution request from the job scheduling means A33 based on the schedule, the job starting means A35 of the host computer H3m.
Has started the execution program A36 of the job requested to be executed.

In each of the embodiments described above, the host computer for re-executing the job when a failure occurs in the host computer is defined by various methods, but these methods are combined in various ways. By holding each priority order in the job management information holding means and the system definition information holding means, the host computer to be re-executed can be appropriately determined.

[0047]

As described above, according to the job re-execution control method of the loosely coupled multiple system of the present invention, when a failure occurs in the host computer that is executing a job, the failure recovery of the host computer is performed. There is an effect that the job can be promptly re-executed by another host computer without waiting.

[Brief description of drawings]

FIG. 1 is a block diagram showing a first embodiment of a job re-execution control method for a loosely coupled multiplex system according to the present invention.

FIG. 2 is a flowchart showing an example of an operation at the time of a job input process of the first embodiment.

FIG. 3 is a flow chart showing an example of operation when a failure occurs in the host computer in the first embodiment.

FIG. 4 is a block diagram showing a second embodiment of the job re-execution control method of the loosely coupled multiplex system according to the present invention.

FIG. 5 is a block diagram showing a third embodiment of the job re-execution control system of the loosely coupled multiplex system according to the present invention.

FIG. 6 is a block diagram showing a fourth embodiment of the job re-execution control method for the loosely coupled multiplex system according to the present invention.

[Explanation of symbols]

A1, A11, A21, A31 Host failure recognition means A2, A12, A22, A32 Job re-execution preparation means A3, A13, A23, A33 Job scheduling means A4, A14, A24, A34 Job management information holding means A5, A15, A25 , A35 Job activation means A6, A16, A26, A36 Execution programs B1, B11, B21, B31 Job control language B2, B12, B22, B32 Job control language translation means C11 Terminal C12 Job execution host registration means D21 User management information storage Means D22 User management information acquisition means E31 Host information acquisition means E32 Job class information acquisition means F31 System definition information holding means G1, G11, G21, G31 Host monitoring devices H1, ~ Hm, ~ Hn, ~, H11, ~ H1m, ~ H1
n, ~, H21, ~ H2m, ~ H2n, ~, H31, ~
H3m, ~ H3n, ~, Host computer

Claims (5)

[Claims] 1. A host monitoring device connected to a plurality of host computers and each of the host computers, and when any one of the host computers fails, the host monitoring device notifies the other host computer of the failure. In a job re-execution control method for a loosely coupled multiplex system, the first host computer includes (A) job management information holding means for holding job control information for a job input in the host computer. (B) When the job control language is input and translated, and the job control language designates the third host computer to re-execute the job when a failure occurs in the second host computer. The job whose registration is registered in the job management information holding means as job control information The language translation means and (C) when the failure occurs, the host monitoring apparatus notifies the second host computer of the failure and recognizes it, thereby referring to the job control information of the job management information holding means. And (D) receiving the job re-execution request from the host failure recognizing means, and thereby executing the job management information holding means. Job re-execution preparation means for making a job re-scheduling request to the third host computer designated by the job control information, and (E) job control from the job control language translation means. Receives information to schedule jobs for execution on multiple host computers Receiving a job re-scheduling request for the third host computer by the job re-execution preparation means and re-scheduling the job to re-execute the job of the second host computer to the third host computer. A job scheduling unit that requests execution, and (F) a job requested to be executed when each of the plurality of host computers has a job execution request based on the schedule from the job scheduling unit. A job re-execution control method for a loosely coupled multiplex system, comprising: 2. A host computer according to claim 1 for re-executing a job when a failure occurs (A) is not specified or the specified specification is changed, When a failure occurs in the second host computer, a terminal for inputting a designation of the third host computer which should re-execute the job, and (B) a failure occurs in the second host computer from the terminal. When the third host computer to re-execute the job is designated, the job re-execution host registration means for registering the designation as new job control information in the job management information holding means is provided. The job re-execution control method for a loosely coupled multiple system according to claim 1, wherein the job re-execution control method is provided for the host computer. 3. (A) User management information storage means for storing the designation of the host computer according to claim 1 for re-executing a job when a failure occurs,
(B) a user management information acquisition unit for acquiring the designation of the host computer for re-executing the job from the user management information storage unit for the user of the faulty job; and (C) the second. If the host computer for re-execution of the job is not designated in the job management information holding means according to claim 1 when a failure occurs in the host computer, the use of the failed job A user to the user management information acquisition unit to acquire the designation of the host computer for the user, register the designation as new job control information in the job management information holding unit, and perform a job failure Host failure recognition means according to claim 1 for making a re-execution request to the first host computer. Claim 1 job re-execution control method of the loosely coupled multiplexing system according to. 4. (A) System definition information holding means for holding the designation of a host computer according to claim 1 for re-executing a job when a fault occurs, and (B) a fault. (C) the second host computer, which acquires the designation of the host computer for re-executing the job from the system definition information holding means for the second host computer in which When a failure occurs,
When the host computer for re-execution of the job is not designated in the job management information holding means according to claim 1, the second host computer is given to the host information acquisition means. By
A method of acquiring a designation of the third host computer for the second host computer, registering the designation as new job control information in the job management information holding means, and requesting re-execution of the failed job. 2. The job re-execution control method for a loosely coupled multiplex system according to claim 1, further comprising: host failure recognition means according to claim 1 for the first host computer. 5. (A) System definition information holding means for holding the designation of the host computer according to claim 1 for re-executing a job when a failure occurs, and (B) Job class information acquisition means for acquiring the designation of the host computer for re-executing the job from the system definition information holding means for the job class of the faulty job; and (C) the second host computer. When a failure occurs, if the host computer for re-execution of the job is not specified in the job management information holding unit according to claim 1, the job class of the failed job is set to the job class. By giving it to the class information acquisition means, the designation of the third host computer for the job class is obtained and the instruction is given. Host fault recognition means according to claim 1, wherein the host failure recognition means is registered as new job control information in the job management information holding means, and a re-execution request is issued to the failed job. 2. The job re-execution control method for the loosely coupled multiple system according to claim 1.