JPH0679301B2 - Job end notification method - Google Patents

Description

DETAILED DESCRIPTION [Outline] In a cooperative process between a host and a workstation in a computer system in which the host and the workstation are connected by a half-duplex communication path, the host asynchronously operates from the workstation. The end status of the job requested to be submitted is managed by a flag, and the status is immediately notified by referring to the flag in response to a job end notification request issued synchronously from the workstation.

[Industrial application field]

The present invention relates to a computer system capable of host / workstation cooperative processing, and particularly, in a computer system in which a host and a workstation are connected by a half-duplex communication path, the host to the workstation Regarding the job end notification method.

[Prior art] Regarding distributed program development In recent years, workstations have achieved remarkable processing capabilities. However, when the workstation and the host are combined, the host can not effectively use the function of the workstation because the workstation can be seen only as an ordinary terminal from the host.

Therefore, attempts have been made to connect the workstation to the host and distribute the processing such as EDIT between the host and the workstation. This is called distributed program development.

Job Execution Server In distributed program development, processing is distributed from the host to the workstations, and the load on the host is reduced by linking the host and the workstation, but on the workstation, a vast memory space is created. Jobs that cannot be executed may occur due to necessity.

For this reason, it was considered to be able to submit jobs from the workstation to the host. On the host, the job execution server fulfills this function.

Conventionally, this job execution server was intended to have the function of performing time-division processing similar to TSS for queued jobs (jobs that are connected to an input queue and executed sequentially). Therefore, the job execution server has the following functions.

-SUBMIT (batch job request) -STATUS (job status check) -CANCEL (job invalidation) However, in developing a distributed program, it is necessary to hide the host system world from the workstation. Is done.

1) In SUBMIT, do not let the user create JCL (job control statements) for starting jobs. The user can automatically create the JCL by assigning parameters (the sensor administrator can see the world of the host, so prepare the necessary catalog procedures for the job).

2) In order to eliminate the need for the workstation side to be aware of the job name of the host, prepare a workstation job name on the workstation side separately from the host job name.

-The workstation job name shall be unique to the job of the workstation user.

-The job execution server names the host job name so that it is unique on the host system.

-Associate the workstation job name with the host job name in the VSAM file of the host system.

Of these, a workstation job management pool (VSAM file) exists to implement the function of 2).

Configuration of Job Execution Server for Distributed Program Development FIG. 4 shows the configuration of a conventional distributed program development system.

In FIG. 4, 31 is a workstation, 32 is a host, 3
3 is manager space, 34 is distributed program development control program, 35 is pool accessor, 36 is workstation job management pool, 37 is server space, 38 is server task, 39 is O
The job management unit of S.

In a distributed program development system, a workstation
When a LOGON request is made from 31, the host 32 is connected to a dedicated manager space 33, like the TSS, where the server task 38 is allocated in the server space 37 for that user (in the case of TSS, Is allocated space). When the server task 38 is assigned to the server space 37, the path is switched between the workstation 31 and the server task 38.

The workstation job management pool 36 manages workstation job names and host job names in association with each other, and is accessed by the pool accessor 35.

The server task 38 will be described in detail. This is a host / workstation distributed (cooperative) processing function having the configuration shown in FIG.
Depending on the request content, the corresponding job execution server
Element servers such as 42, file transfer server 43, and translation server 44 are selected and activated.

The job execution server 42 is as described above, and performs control such as requesting the job management unit 39 to submit a job requested by the workstation 31. The file transfer server 43 controls file transfer between the host and the workstation. The translation server 44 performs language translation processing. Hereinafter, any server can be provided.

[Problems to be solved by the invention]

In the conventional distributed program development system shown in FIGS. 4 and 5, the communication between the job execution server 42 of the host 32 and the workstation 31 is performed by the half-duplex communication protocol, and the server task 38 Is limited to the fact that a session exists until it is disconnected, and that only one response can be returned per request.

The host must notify the workstation of the end of the job within these restrictions.

In this case, when a job submission request is sent from the workstation to the host server task, the server task performs two things, job submission / completion monitoring, and notifies the workstation when the job is completed. There must be. Without such a synchronous method, the event could not be notified when the job ended.

However, in this synchronous method, since it is half-duplex communication, there is a problem that the protocol for operating the host cannot be sent from the workstation until the end of the job is notified after the job submission request. there were. This problem is very inconvenient for a job that requires a long time to complete, such as a queued job.

[Means for solving problems]

In order to solve the above-mentioned conventional problems, the present invention uses a job end notification as a function different from a job submission request. First, the workstation issues a job submission request to the host, and the workstation responds to the submission. receive.

At this time, the host requests the job management unit to submit a job and creates an entry in the workstation job management pool to store information such as the job name of the submitted job.

After that, on the host side, the job requested to be submitted starts,
When the job is completed, the fact is reflected in the contents (flag indicating completion) of the workstation job management pool. In this way, the workstation job management pool always reflects whether or not a job has finished.

In the workstation, after submitting a job, a timer is used to periodically issue a job end notification request.

At the host, when an end notification request is received from the workstation, the server task refers to the contents of the workstation job management pool (a flag indicating the end) and notifies the workstation whether the job has ended or not. To do.

On the workstation, select the information of "Job finished" or "Job not finished" to be returned in response to the end notification request,
If it is “job end”, the job end notification request that is regularly performed is ended, and the user is notified that the job has ended.

In this way, by dividing the period from the job submission request to the job end notification into each function, shortening the time required for each request-response, and making each independent, the workstation submits the job. After making a request, it is possible to perform an operation such as inputting another job without waiting for the end of the job.

FIG. 1 shows the basic configuration of the present invention.

In FIG. 1, 1 is a workstation, 2 is a job execution function, 3 is a host, 4 is a job execution server, 5 is a job submission function, 6 is a job end notification function, 7 is a job management unit, 8 is a workstation job It is a management pool.

The workstation 1 has a job execution function 2 for performing host / workstation cooperation in processing tasks such as distributed program development.

The job execution function 2 includes a job submission request function for the host 3, a job completion notification request function for periodically requesting notification of the job completion status to the host 3 using a timer, and a job completion notification from the host 3. It also has a job end determination function that determines the end of a job based on the status.

The job execution server 4 of the host 3 is a host-side control function for host / workstation cooperation processing, and has a job submission function 5 and a job end notification function 6.

Upon receiving a job submission request from the workstation 1, the job submission function 5 notifies the job management unit 7 which is a part of the OS to submit (start) the requested job. At the same time, an entry for the submitted job is created in the workstation job management pool 8. The entry is composed of a key of the job name and a flag indicating the end status of the job.

The workstation job management pool 8 displays a flag for the status of each job requested from a plurality of workstations or jobs requested from a plurality of tasks in one workstation when the job is completed. .

When a job end notification request is issued from the workstation 1, the job end notification function 6 reads an entry with the corresponding job name in the workstation job management pool 8 as a key, and according to the display of the flag, the job is notified to the workstation 1. Notify the end.

[Action]

The operation based on the basic configuration of the present invention shown in FIG. 1 will be described with reference to the processing example shown in FIG. The constituent elements in FIG. 2 are the same as those in FIG. 1, but the job end notification functions 6, 6 ′, 6 ″ represent those at different operating points in sequence.

The job execution function 2 of the workstation 1 enters a waiting state with a timer after the job submission request is completed, and outputs a job completion notification request after the timer is completed. as a result,
If the job is completed, the user is notified that the job is completed. If it has not finished, it will wait again and output a termination notification request.

The job submission function 5 of the host 3 requests the job management unit 7 to submit a job, then sends back the request status to the workstation 1, and at the same time creates an entry in the workstation job management pool 8 using the job name as a key. To do.

When a job end notification request is input, the job end notification function 6 looks at the contents of the entry for the job name in the workstation job management pool 8 and indicates that the job has ended (the flag is ON. If so, notify the end. If it does not indicate that it has finished, notify that it has not finished.

The workstation job management pool 8 turns on the end flag of the entry of the job name at the end of the job. The job execution function 2 of the workstation 1 periodically outputs a job end notification request at a cycle determined by a timer until the job ends. The contents of the workstation job management pool 8 are updated at the timing of job termination. The job end notification function 6 is called every time a request is made, but the job submission function 5 is called only once.

With such a flow, it is possible to notify the work station 1 that the submitted job has finished at about the same timing as the end time.

〔Example〕

FIG. 3 shows a system configuration on the host side according to an embodiment of the present invention.

In FIG. 3, 10 is a common area also called a space mutual communication area (CSA), 11, 12, 13 are server spaces, 14 is a job management space, 15 is a manager space, 16a, 16b, 16c are server spaces 1
Three job execution servers assigned to 1, 17a, 17b, 17
c is three job execution servers assigned to the server space 12, 18a, 18b, 18c, and 18d are four job execution servers assigned to the server space 13, 19 is a job management routine, and 20 is
WTO exit, 21 is chain task, 22 is pool accessor,
Reference numeral 23 represents a workstation job management pool.

A plurality of job execution servers can be assigned to each of the server spaces 11, 12, and 13.

Server space 11, 12, 13, job management space 14, manager space
Data communication between 15 is performed via the common area 10.

When the job execution server 4 executes the job requested to be submitted to the job management routine 19 and ends, WTO (Write To O
A termination message is output from the perator) exit 20 to the console (not shown).

The WTO exit 20 detects a job end message, creates job end data, and passes it to the manager space 15 via the common area 20.

The rechaining task 21 of the manager space 15 submits the job name (correspondence between the workstation job name and the host job name) of the job requested by each of the job execution servers 16a to 18d and the job status. When they are connected to the chain in sequence and the job disappears, they are removed from the chain and passed to the pool accessor 22. The pool accessor 22 manages this with the workstation job management pool 23.

Upon receiving the job end data from the WTO exit 20, the pool accessor 22 accesses the workstation job management pool 23 and updates the job status of the corresponding job name to the end display.

In response to a reference request from the job execution server, such as when a job end notification request is received from a workstation, the pool accessor 22 refers to the workstation job management pool 23, reads the job status of the corresponding job name, and requests it. Notify the original job execution server.

〔The invention's effect〕

According to the present invention, in the host / workstation cooperative processing of a computer system in which the host and the workstation are connected by a half-duplex communication path, a job end notification is sent from the host to the workstation without a large delay. Moreover, the operation of the workstation that has submitted the job can be performed without being held in a waiting state, and the processing efficiency of the system can be improved.

[Brief description of drawings]

FIG. 1 is a block diagram of the principle of the present invention, FIG. 2 is an explanatory diagram of a processing example showing the operation of the present invention, FIG. 3 is a block diagram of an embodiment of the present invention, and FIG. 4 is a conventional host workstation. FIG. 5 is a block diagram of the distributed program development system by cooperation, and FIG. 5 is a block diagram of the server task in FIG. In Figure 1, 1: workstation 2: job execution function 3: host 4: job execution server 5: job submission function 6: job end notification function 7: job management section 8: workstation job management pool

Claims (1)

[Claims] 1. In a computer system in which a host and a workstation are connected by a half-duplex communication path and host / workstation cooperation processing is performed, the host is a job for a job submitted by the workstation. Workstation job management pool (8) that holds the name and the flag indicating the status of the job in association with each other
When a job end is detected, the job end status display is set in the corresponding job name flag in the workstation job management pool (8), and a job end notification request issued from the workstation is received. At this time, the flag of the corresponding job name in the workstation job management pool (8) is checked to notify the workstation of the completion of the job, and the workstation uses the timer after requesting the host to submit the job. A job end notification method characterized by periodically issuing a job end notification request to the host.