JPS608946A - Halt processing system of job - Google Patents

Abstract

PURPOSE:To save the result of a job being executed, in a storage device by providing a table for always updating and holding the name of a job being executed by a computer, on the storage device, and reading out the job name from the table, when halt processing of the job is instructed. CONSTITUTION:When the fault of a computer 1 is informed to an operator through a console 8, the execution of a main job is transferred to an operating system OS91 through a console 9, and the command requiring processing to a sub-job is given. The OS91 executes a holding processing of sub-jobs 24, 25 and 26 waiting for the execution of a table 47, subsequently, executs halt processing of sub-jobs 21, 22 and 23 being in the course of execution of a table 46, and writes the contents of the related storage device in a fault file. Next, I/Os 3, 5 are read out of a table 41, and made an I/O which is capable of access to the OS91. Subsequently, the name of a main job which is executed by the computer 1 is read out of a table 43, a job queue 30-1 is accessed, and main jobs 11 and 12 are restarted.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to a process for interrupting a job in progress.

[Background of the invention]

In recent years, information processing systems, especially online systems, require high reliability, so system configurations using multiple computers have been adopted. For example, when a failure occurs in the first computer, the 5G2 computer Many systems are designed to back up a first computer and continue the job that was running on this computer.

In this case, the second computer 18 can simply stand by as a spare computer, but it can be used productively by executing batch jobs while the first computer is operating normally. There are many systems that are configured to do so. In such a system, when a failure occurs in the first computer, the second computer stops the job that was being executed on this computer (hereinafter referred to as the slave job) and restarts the job on the first computer. Normally, the execution of the currently executing job (hereinafter referred to as the main job) is taken over and continued. In this case, there is a problem that whether the restart of the main job is quick or the subordinate job that is being executed -1 is stopped, so it takes a long time to start the subordinate job.

One way to solve this problem is to interrupt the execution of the subordinate job that is currently being executed through operator intervention, and to quickly restart the execution later, the related storage device's self-response (mid-execution result) is displayed. There is a way to ensure that the process is interrupted, such as by dumping it to a file. However, this method
There is a problem in that the operator's intervention time is required, and the restart start-up of the main job is delayed.

The above problem will be specifically explained below with reference to the drawings. FIG. 1 is an example of a configuration diagram of a system composed of a plurality of X machines in total. Computer 1 is a computer that executes the main job, Computer 2 is a computer that executes subordinate jobs, Console 8 is a console device that is connected to Computer 1 and controls and monitors this computer 1, and Console 9 is connected to Total/X Iso 2. A console device and an input/output device that control and monitor this computer 2 [(Ilo) 3, 4, 5, and 6 are input/output devices that are connected to the computer 1 and the computer 2 and are shared by the two-hand calculator. . ossl is a physical source such as a computer 1 and its attached devices, and a file 7j.
ti - Operator Ink System or Monitoring Program for Intelligent Entities, 0891 is such an operator ink system or monitoring program for a total of $12. l103-6 are direct access storage devices as kings, ■/G3 stores job queue 30-1 (used by JS81), and 1104 stores 0891
A job queue 30-2 used by is stored therein.

FIG. 2 schematically shows the order of the job queue 1-30-2. In the job queue, jobs currently being executed and jobs waiting for execution are classified into job classes 3O-2A and 3O-2.
It is grouped by 2B8 and 3O-2C and registered as an input job queue.

The slave jobs 21, 22 and 23 are running slave jobs, and the slave jobs 24, 25 and 26 are slave jobs waiting to be executed, and each contains control information such as job name used for job schedule links.

With the above configuration, monitor console 9 as follows: 3
- Explain the operator's intervention situation and the operation of 0891. First, when the operator inputs a command (instruction) requesting the display of subordinate jobs waiting to be executed from the console 9, 0891 searches for job queue 30=2 and displays subordinate jobs 24.25g and 26 waiting to be executed on the console 9. do. Next, the operator selects slave jobs 24, 255 and 2.
0891 enters a command requesting to hold or cancel job queue 3 one by one.
0-2 and place these slave jobs one by one on hold or cancel them. Next, when the operator inputs a command to display the subordinate jobs that are being executed, 0891 accesses the job queue 30-2 and displays the subordinate jobs 21, 22 and 23 that are being executed on the console 9.
to be displayed. Next, when the operator inputs a command requesting to suspend the slave jobs 21, 22 and 23 one by one, the 0891 performs the above-described suspension process for these slave jobs one by one.

As described above, according to the conventional method, it takes a long time to process the suspension of the slave job for the following reasons, resulting in the problem of delaying the start-up of the main job.

(1) The number of times the operator inputs commands is large, so it takes time for operator intervention.

(2) It takes time for 0891 to access the job queue 30-2 and retrieve control information regarding a subordinate job that meets the conditions.

[Purpose of the invention]

An object of the present invention is to eliminate the above-mentioned problems and to speed up job interruption processing.

[Summary of the invention]

The present invention provides a computing system that performs job schedule linking based on a job queue, and the computing system stores a table that constantly updates and holds job names of jobs being executed on a computer, separately from the job queue. The method is characterized by a job interruption processing method that is provided on the device and that, when an instruction to interrupt a job being executed is given from the console, reads the job name from the table and saves the mid-execution result of the job to a storage device. Particular [Embodiment of the Invention] An embodiment of the present invention will be described below with reference to the drawings.

The target system configuration is shown in FIG.

However, jobs 103 to 6 are input/output devices that are connected to total It is assumed that only 103 and 105 are accessed @, and 0881 is accessed other than 103 and 105. Like Job 2 and Saitama? )891 can only access l104.5 and l106, and (JS91 can access I/r5 other than these).

FIG. 3 is a diagram showing the configuration 8 and P responses of the management table 40 according to the present invention. It is assumed that this management table 40 is stored on l105 in the T configuration diagram shown in FIG. The table 41 records Ilo that is logically connected to the computer 1, and in this example, Ilo
03 and l105 are connected to computer 1. Table 42 is a table in which king job names are recorded,
It consists of a table 43 and a table 44. table 43
is a table that stores the names of the main jobs being executed. This example shows that main job 11 and main job 12 are being executed. Table 44 is a table in which names of main jobs waiting in the middle row are stored. This example shows that main jobs 13, 142, and 15 are waiting for execution. Table 45 is a table in which subordinate job names are recorded, and consists of table 46 and table 47. The table 46 is a table in which names of subordinate jobs that are being executed are stored, and in this example, indicates that subordinate jobs 21, 22, and 23 are being executed. table 4
7 is a table in which names of subordinate jobs waiting to be executed are stored, and in this example, subordinate jobs 24, 25, and 26 are shown waiting to be executed.

The operation of one embodiment of the present invention will be explained below.
1 is I/? connected to computer 1. ) is changed, the contents of the table 41 are rewritten. Also, 0881 is
The contents of the table 42 are rewritten each time the status of the main job being executed or waiting to be executed under the control of the 0881 changes. Similarly, 0891 rewrites the contents of table 45 every time the status of a subordinate job that is being executed or is waiting to be executed changes under the control of 0891.

Next, when a failure occurs in computer 1, the main job being executed on computer 1 becomes unprocessable. Faults in the machine 1 are reported to the operator through the console 8 or through other monitoring mechanisms not shown. Here, the operator inputs a command to 0891 through the console 9 requesting to take over the execution of the main job and request processing for the subordinate job. 0891 receives this command (Fig. 4, 71) cσ891 first reads table 4 on work/65.
7, the names of subordinate jobs waiting to be executed are stored, and the pending processing is performed for these subordinate jobs (72 in the same figure).

In other words, the job queue 30-2 is accessed and job 2 is
4.Place or cancel 25 and 26. Next, the names of the subordinate jobs being executed are read from the table 46, and interruption processing is performed for these subordinate jobs (73 in the same figure). That is, the contents of the storage device associated with slave jobs 21, 22 and 8 and 23 are dumped to the failure file in order to speed up the restart later on. Next, 0891 is the work 10 number logically connected to calculation a1 from table 41, that is, l
10a and I 705, and
is added as Ilo that can be accessed by 08H (74 in the same figure). Engineering 105 is already here! rt is connected to that computer, so in this case, is the new work 103? )89
It will be added as Ilo under 1 control. then 0
891 reads the name of the main job being executed on computer 1 from table 43, accesses job queue 30-1, and restarts main jobs 11 and 12 (see 75 in the same figure).
).

Since the method of restarting the main jobs 11 and 12 falls within the scope of the prior art, a detailed description thereof will be omitted. Note that since the table 44 stores the names of main jobs waiting to be executed, the 0891 may display these names on the console 9 as an overload. In the above embodiment, 0891 performs a hold process on the slave jobs waiting to be executed on the table 47, but this is not necessarily necessary and may not be performed depending on the system.

Alternatively, only the job being executed on computer 2 may be interrupted without computer 1. In this case, the management table 40
There is only table 45 on top, so it is definitely l105.
It is not necessary to have the table 45 on the input/output device, but the table 45 may be placed on the main storage device of jiLa2.

As described above, according to the above-mentioned embodiment, the operator only has to perform the first command manually, and processes the main job that takes over execution, the subordinate jobs that are currently running, and the subordinate jobs that are waiting to be executed one by one. Operator intervention is reduced because there is no need to enter commands to request Also, in addition to the job queue 30-1 and the job queue 30-2, the main job names and slave job names related to those currently being executed and those waiting for execution are held as a management table 40 on the machine 105. 30-
The number of accesses of 2 is reduced.

〔Effect of the invention〕

According to the present invention, there is an effect of speeding up job interruption processing.

[Brief explanation of the drawing]

Fig. 1 is a configuration diagram showing an example of the configuration of the needle valve system, Fig. 2 is a diagram showing the contents of the job queue, Fig. 3 is a diagram showing the structure and contents of the management table, and Fig. 4 is a diagram showing the processing hand 711 of 0891.
1ij is a flowchart showing the process. 1.2... Computer 3.4, 5.6... l10 8.9... Console 21-26... Sub job 30-1. 30-2... Job queue 40... Management Tables 41-47...Table agent Patent attorney Akira Takayoko / Hidai 22

Claims (1)

[Claims] 1. It has a console and a storage device that are connected to a total of SM and a total of 31L machines, and has a job queue that is a queue for input jobs.
In a computing system that performs job ψ schedule linking based on a queue, the computing system
In addition to the queue, a table is provided on the storage device that constantly updates and retains the job names of jobs being executed on the computer, and the console or the computer instructs the computer to suspend processing of the jobs being executed on the computer. 8. A job interruption processing method, characterized in that, when the job is terminated, the job name is read from the table and the mid-execution result of the job is saved in the storage device.