JPS6197750A - Restart system for on-line job - Google Patents

Abstract

PURPOSE:To restart early an on-line job (specific job) by supplying with priority the on-line job before the restoration is started for a general job staying within a system in a pre-operation system and in a system rerise mode after the system operation is stopped by a fault and then performing various types of restoration proper to the specific job together with the restoring action of the general job. CONSTITUTION:When a system has a rise, a main memory, etc. (not shown here) within the system are initialized by a system starting part 1. This initialization is informed to an execution waiting job restoring part 2 as well as to a specific job starting part 9. The part 2 retrieves an execution waiting job out of a jot state memory part 6 and registers it to a job control memory part 13. In the same way, both a job waiting for output and a job under execution are registered to the part 13 as the output waiting jobs as well as to an output queue memory part 8. While the part 9 supplies an on-line job as a specific job and registers it to a job queue memory part 11. Then a specific job executing part 12 performs various types of restoration proper to the specific job. Thus an on-line job is restarted.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention provides an online job restart method at the time of system startup after a computer system is stopped due to a failure [Prior Art] In a conventional computer system, the system stops due to a failure. When the system is restarted later, the jobs that were in the execution waiting state during the previous operation are restored to the execution waiting state again, the jobs that were in the output waiting state during the previous operation are restored to the output waiting state again, and the jobs that were in the execution state during the previous operation are restored to the output waiting state again. After restoring jobs that have output information to the output waiting state, it becomes possible to input and schedule jobs, so depending on the number and status of jobs at the time of the previous operation, the startup of jobs after restarting the system may be delayed. In other words, there was a drawback that the resumption of online operations was delayed.

[Problem that the invention seeks to solve]

SUMMARY OF THE INVENTION An object of the present invention is to provide an online job restart method for quickly restarting online operations when the system is restarted after it has been stopped due to a failure.

[Means for solving problems]

According to the present invention, when the system is restarted after the system is stopped due to a failure, online jobs are input preferentially before starting the restoration of general jobs that were stuck in the system during the previous operation, and the online jobs are input in parallel with the restoration operation of the general jobs. This system performs the restoration of each error specific to a specific job, restarts online work, and notifies the online job when the restoration of the general job is completed.

That is, the online job starting method of the present invention is a method for starting an online job that operates in a large-scale on-two-in-one system.
means for inputting and scheduling jobs that were stuck in the system during a previous operation on a priority basis as a specific job, distinguishing them from general jobs, when the system is restarted after the system has been stopped due to damage; In parallel with the execution of the above, means for restoring the job waiting to be executed during the previous operation to the waiting state again, means for restoring the job waiting for output during the previous operation to the output waiting state again, and execution during the previous operation. means for restoring a job in the state having output information to the output waiting state, and means for notifying the special foot '9's job that restoration of the job to the execution waiting state and the output waiting state has been completed. Equipped with

〔Example〕

Embodiments of the present invention will be described with reference to the drawings.

Figure N41 is a block diagram of a system showing an embodiment of the online job starting method of the present invention.

This system 100 includes a system startup unit 1 that starts the system, a job restoration unit 2 that restores jobs that were waiting to be executed during a previous operation to an execution waiting state, and a job restoration unit 2 that restores jobs that were waiting to be output during a previous operation. An output waiting job restoring section 3 restores the job to the output waiting state, a running job restoring section 4 restores the jobs that were in the running state at the time of the previous operation and have blade information to the output waiting state, and a message indicating that all the restorations have been completed. A restoration completion notification unit 5 that notifies a specific job, a job status storage unit 6 that stores the job status at the time of the previous operation, an output request registration unit 7 that registers an output request in the system VC, and an output unit that stores the output request. A queue storage unit 8, a specific job activation unit 9 that inputs and schedules online jobs as specific jobs, a specific job activation information storage unit 10 that stores information for starting specific jobs, and a job queue that stores job schedules. It is composed of a storage section 11, a specific job execution section 12 in which an online program operates, and a job management storage section 13.

Next, the operation of this embodiment will be explained. When the system is started, the system startup section IK initializes the main storage device (not shown) in the system, and notifies the waiting job restoration section 2 and the specific job startup section 9 of the initialization. The waiting job restoring unit 2 searches the job status storage unit 6 for a job waiting to be executed during the previous operation, and registers the job in the job management storage unit 13 as a job waiting to be executed. When the restoration by the execution waiting job restoration unit 2 is completed, the output waiting job restoration unit 3 is called.

Similar to the execution waiting job restoration section 2, the output waiting job restoration section 3 searches the job status storage section 6 for the output waiting job from the previous operation, registers it in the job management storage section 13 as an output waiting job, and at the same time requests the output. The output request message is registered in the output queue storage unit 8 using the registration unit 7. When the output waiting job restoring unit 3 completes the restoration, the running job restoring unit 4 is called. The running job restoration unit 4 searches the job status storage unit 6 for jobs that have internal output information of jobs in the execution status during the previous operation, in the same way as the execution waiting job restoration unit 2 and the output waiting job restoration unit 3, and stores them in the job management storage. Part 13
is registered as an output waiting job in the output request registration section 7.
The output request message is registered in the output queue storage section 8 using .

On the other hand, the specific job starting unit 9 inputs an online job registered in advance in the specific job starting information storage unit 10 as a specific job, and registers it in the job queue storage unit 11. A specific job selected by a predetermined scheduling algorithm from the job queue storage unit 11 is subjected to various types of restoration unique to the specific job by the specific job execution unit 12, and online operations are resumed. Thereafter, the specific job execution unit 12 receives a restoration completion notification for the job during the previous operation from the restoration completion notification unit 5.

In this way, online operations can be restarted quickly by performing job restoration from the previous operation and various restorations of specific jobs in parallel.

〔Effect of the invention〕

As described above, the present invention has the effect that online operations can be restarted early by starting specific jobs preferentially when the system is restarted.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing one embodiment of the present invention. DESCRIPTION OF SYMBOLS 1... System startup part, 2... Execution waiting job restoring part, 3... Output waiting job restoring part, 4... Running job restoring part, 5... Restoration completion notification part, 6 ... job status storage section, 7... output request registration section, 8... output queue storage section, 9... specific job starting section, 10... specific job starting information storage section, 11... Job queue storage unit, 12--Specific job execution unit, 13--Job management storage unit.

Claims (1)

[Claims] When restarting an online job running on a large-scale online system after the system has stopped due to a failure,
A means for preferentially inputting and scheduling a job that was stuck in the system during a previous operation as a specific job to distinguish it from a general job, before starting restoration, and waiting for execution during the previous operation in parallel with the execution of the specific job. means for restoring a job in an execution state again to an execution waiting state; means for restoring a job in an output waiting state during the previous operation to an output waiting state; and a means for restoring a job in an execution state during the previous operation to an output waiting state. 1. An online job restart method comprising: means for restoring the specific job to the execution wait state and output wait state; and means for notifying the specific job that the restoration of the job has been completed in the execution wait state and the output wait state.