JPS593609A - Work station switching control system for fault occurrence - Google Patents

Abstract

PURPOSE:To run jobs continuously by disconnecting all jobs under the control of a work station which becomes unused and concatenating them to jobs under the control of another usable work station. CONSTITUTION:If an unrecoverable error is detected during a job for access to a work station WS2, a message regarding the error occurrence is reported to a work station WS1 that a center operator attends. At the same time, the PDCB and ARCB of a job control block 10 are disconnected from the control blocks WDCB and MUCB of the work station WS2 and the job control block 10 is put in a stand-by state for connection with another usable work station. The center operator activates a free work station WS3 and generates a command for switching jobs from the WS2 to the WS3 to restart running the jobs.

Description

DETAILED DESCRIPTION OF THE INVENTION Technical Field of the Invention The present invention relates to a switching control method for a display terminal device, that is, a workstation, in a data processing system having a plurality of display terminal devices, and in particular to a method for controlling the switching of display terminal devices, that is, workstations, in a data processing system having a plurality of display terminal devices. If a job is unable to continue processing due to an unrecoverable hard error on that workstation, most of the resources it was using at that time are put on hold without abnormally terminating the job. The present invention relates to a workstation switching control scheme that allows takeover for re-execution under other available workstations.

Background of the Technology In the conventional job execution method, as shown in Figure 1, a demand job DJ running in an interactive manner at a workstation and a background job BJ such as printer output processing that occurs as a derivative job are executed at a certain workstation. If there is a job that collectively manages these jobs (this is called a management job), an unrecoverable error occurs when this management job executes read or write processing to the workstation, for example. When a transmission error due to a hardware failure is encountered, the management job sends demand job DJ and pack ground job B.
After I abnormally terminated everything in J, I had to abnormally terminate myself.

FIG. 2 is a system configuration diagram of a conventional workstation control method. In the figure, 1 and 2°3 are workstations WSI, WS2°WS3.4 are hosts, 5 is an MSP processing unit, 6 is an MCP processing unit, and 7 is a WDC.
B, 8 represents a MUCB, and 9 represents a job management block.

Note that WSl is the center operator workstation,
WS2. It is assumed that WS3 is a workstation that is processing a job.

The MSP processing unit 5 uses MSP (mask schedule program) to analyze messages sent from the workstation, calls necessary applications (jobs), and performs necessary control for execution.

The MCP processing unit 6 controls message communication with the workstation using an MCP (message control program).

WDCB7 and MUCB8 are workstation W
It constitutes the management block of S2.

The WDCB (workstation device control block) 7 is a control block that is set corresponding to the workstation when the MSP processing unit 5 identifies the "sign on" information input by the operator from the workstation. Workstation and job processing execution mechanism (TCB).

It holds a pointer indicating correspondence with MURB, MUCB') and other management information such as workstation machine number and line number.

A MUCB (message unit control block) 8 is connected to the WDCB and manages message communication requests for each job to the workstation.
ARCB manages the memory area used for communication, which will be described later.
used to call (point to).

The job management block 9 is a general term for a group of blocks used for job execution management, including TCB (task control block), JCB (job control block), and MURB (message unit). request block), ARCB (access block),
request control block).

The TCB holds control information and task status necessary for executing tasks that process jobs. The JCB holds management information such as job status and priority. MURB is used to call ARCB when a job requires the transmission of a message to a WS. ARCB is
A memory area carved out by the MCP is managed for each message transmission request.

The system described above manages the execution control information of the job management block 9 through the WDCB of the workstation management block generated uniquely to each WS by the MSP, so for example, if WS2 is disconnected due to a failure,
Unless the job management block 9 becomes referenceable, execution cannot be taken over by another workstation, and as described above, all the jobs being executed have to be abnormally terminated.

Therefore, in order to restart the process from another available workstation, the workstation operator repeats the startup operation of resubmitting all of the demand job DJ and packground job BJ that were previously terminated abnormally. There was a problem in that the process had to be returned, which delayed the process and increased the burden on the operator.

Object and Structure of the Invention The object of the invention is to disconnect all jobs that were running under an unusable workstation, connect them to another available workstation, and then disconnect all jobs that were running under that workstation from a workstation that has become unusable. The objective is to provide a means to allow job processing to continue.

As a solution to this problem, the present invention does not directly connect and manage jobs to a workstation management unit (WDCB) as in the conventional method, but temporarily manages jobs under operator attributes. By connecting a workstation through a means for managing operator attributes, even if a workstation that has become unusable is disconnected from a job (and therefore from operator attributes), the means for managing operator attributes can be referenced. and then use it again as another available workstation management method (W
DCB), it is possible to continue executing jobs in progress regardless of changes in workstations.

As a configuration for this purpose, the present invention provides a workstation management means provided for each activated workstation in order to control communication to the activated workstation in a data processing system consisting of a host and a plurality of workstations. A job management means provided for each job processed under one workstation, and an operator attribute management means for combining the job management means with the workstation management means (operator attribute management means for combining the job management means with the workstation management means) are provided in the host. When an irrecoverable failure occurs in access from a host to a station, the connection between the workstation management means and operator attribute management means for the workstation is severed, and job processing is interrupted by placing the job management means on hold. On the other hand, a vacant workstation is newly activated and a corresponding workstation management means is provided, and the operator attribute management means separated from the workstation management means of the faulty workstation is
By coupling to the workstation management means of the newly activated workstation, the job management means placed in the suspended state becomes functional under the new workstation, and the execution of the interrupted job processing is resumed. Embodiments of the Invention The present invention will be described below with reference to embodiments.

FIG. 3 is a block diagram of an embodiment of the workstation control system according to the present invention, and FIG. 4 is an explanatory diagram of its operating procedure.

FIG. 3 is largely based on the conventional workstation control scheme shown in FIG. 2, with like reference numbers indicating the same nine components. In the figure, 1.2.3 are workstations WSI and W, respectively.
S2. It is WS3. Also, 5 is the MSP processing unit, and 6 is the MC
P processing unit, 7 is the WDCB of WS2, 8 is the same<MUCB of WS2, 10 is the job management block improved by the present invention, 11 is the PDCB, 12 is the WDCB of WS3, 1
3 represents the same MUCB of <WS3. However, in WS3, initially, Affi, WDCB, and MUCB are not set in an empty state.

TCB and MURB in job management block 10
is pointed by a PDCB (7' Low-Fail Device Control Block) 11 inserted according to the present invention, which is different from the case of the conventional system shown in FIG. This PDCBII is pointed to by WDCB7.

Therefore, in the conventional system, WDCB directly points to TCB and MURB, whereas in this embodiment, WDCB points to TCB and MURB via PDCB.
There is a difference in the points to be made. However, in terms of normal operating functions, both are exactly the same.

As shown in FIG. 4, when an operator signs on, the MSP processing unit provides a PD for that operator.
Get a CB. The PDCB holds attributes such as the authority of the operator.

Next, workstation switching control will be explained. When a job accessing the workstation WS2 detects an unrecoverable error, a message regarding the occurrence of the error is sent to the workstation WS1 where the center operator is located. At the same time, the job management blocks PDCB and ARCB are
It is separated from the management blocks WDCB and MUCB of workstation WS2, and the job management block 1o is placed in a waiting state for connection to another available workstation.

Upon receiving the error message, the center operator activates the workstation WS3, which no one is using, and acquires the workstation management blocks WDCB12 and MUCB13fi for WS3. At the same time, a command is input to switch the job of WS2 to WS3, and the PDCBll of the job management block 10 is controlled to be connected to the WD CB 12 as belonging to the operator of WS3.

As a result, job execution can be restarted in WS3 while keeping most of the resources used in WS2 on hold.

FIG. 4 shows the MSP processing unit 5 in the system of this embodiment.
, MCP processing unit 6 and system management unit (not shown)
Regarding the operation functions of O8), the flowchart shows the interrelationship among initialization, system job execution, interactive user job (demand job) execution, and checkground job execution.

Effects of the Invention According to the present invention, job processing can be continued from a workstation that has become unusable without abnormally terminating the job that was running under that workstation. Resubmit the job by
The burden of recreating some created work is eliminated, and processing delays due to failures can be kept to a short time.

[Brief explanation of the drawing]

Figure 1 is a diagram explaining the abnormal termination of an executed job on a workstation, Figure 2 is a system configuration diagram of a conventional workstation control method, Figure 3 is a system configuration diagram of a workstation control method in this embodiment, and Figure 4 is the third
FIG. 4 is an operational flowchart of the system shown in the figure. In the figure, 1 to 3 are workstations, 5 is an MSP processing unit, 6 is an MCP processing unit, 7 and 12 are WDCBs, 8 and 13 are MUCBs, 10 is a job management block, 11
indicates PDCB. Patent applicant Fujitsu Ltd. Representative Patent Attorney Fumihiro Hasegawa (1 other person) No. 111'4

Claims (1)

[Claims] In a data processing system comprising a host and a plurality of workstations, a workstation management means provided for each activated workstation for controlling communication to the activated workstation, and a workstation management means provided under one workstation. A job management means that can be installed for each job processed by the host, and an operator attribute management means for connecting the job management means to the workstation management means are provided in the host. When an unrecoverable failure occurs, the connection between the workstation management means and the operator attribute management means for the workstation is severed, the job management means is put on hold and job processing is interrupted, and the A workstation is newly activated and a corresponding workstation management means is provided, and the operator attribute management means separated from the workstation management means of the faulty workstation is transferred to the newly activated Sareta workstation. By being coupled to the workstation management means, the job management means placed in a suspended state is activated under a new workstation, and the job management means is controlled to resume execution of the interrupted job processing. A workstation switching control method in the event of a failure.