JP2641919B2 - Job schedule stop device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to a multi-host system including one global host and a plurality of local hosts.
The present invention relates to a device for stopping a job schedule of a global job input from a global host and executed on a local host, and particularly to a job schedule stop for stopping a job schedule of a global job accessing the resource in the global host when a failure occurs in the resource in the local host. Equipment related.

〔Overview〕

According to the present invention, in a job schedule stop method, a correspondence defined between a plurality of resources in a local host and a local job class whose schedule should be stopped for its failure is also reflected in a global job schedule. This eliminates the need to define in the global host the association between the resources in the local host and the global job classes whose schedules should be stopped for the faults, thereby simplifying the system definition.

[Conventional technology]

The multi-host system includes one global host for inputting and scheduling global jobs, and one or more local hosts for inputting local jobs and executing local jobs and global jobs. Global hosts also often overlap local hosts.

When some resources become unavailable due to a failure in the local host, it is necessary to stop the schedule of the job using the resource regardless of the global job and the local job. Which job to stop is determined based on the job class declared for each job.

Conventionally, in the job scheduling method, a job class to be stopped in relation to a failure is determined on both the global host and the local host, and the schedule for the job class is stopped, so that the job using the resource is stopped. Had stopped running.

[Problems to be solved by the invention]

However, in such a conventional job schedule stop method, the global job is scheduled on the global host, and the local job is scheduled on the local host. This has to be performed on both the host and the local host, and the system definition is complicated.

SUMMARY OF THE INVENTION The present invention solves the above-described drawbacks, and it is not necessary to define on a global host a correspondence between a resource in a local host and a job class whose schedule should be stopped for its failure, and a job whose system definition is easy The purpose is to provide a schedule stop method.

[Means for Solving the Problems] The present invention is a job schedule stop device in a global host of a multi-host system composed of one global host and a plurality of local hosts,
The local host has a first definition table storing a plurality of job class names declared for a plurality of jobs executed on the local host and a plurality of resources used for the jobs, and a first definition table for the first definition table. First scheduling means for requesting sequential job execution based on a table, and first setting means for setting information for stopping use of a resource in the host in the event of a resource failure in the first definition table in the first definition table Including, the global host is a second definition table that stores a plurality of job class names declared for a plurality of jobs to be executed by a plurality of local hosts and local host names corresponding to those job class names, Based on the second definition table, the global host of the multi-host system including the second schedule means for requesting the corresponding local host to execute the job. In the job schedule stopping apparatus, the first setting means of the local host includes means for notifying the global host of a job class name of a job using the resource when a resource in the local host fails, the global host Includes a second setting unit that sets information for stopping the execution request of the job in the second definition table for the corresponding notification source local host based on the job class name from the notifying unit. It is characterized by the following.

[Action]

In the local host, when a failure is detected in a resource in the local host, control information for stopping a job class schedule related to the failed resource is set in the first definition table by the first setting means, and the failure is set in the global host. Notify the job class name related to the resource. In the global host, the second setting means sets, in the second definition table, control information for stopping the schedule for the notification source local host of the corresponding job class based on the notification. According to the above operation, it is not necessary to define the association between the resource in the local host and the global job whose schedule should be stopped for the fault in the global host.

〔Example〕

Embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of a job schedule stopping device according to an embodiment of the present invention. In FIG. 1, the job schedule stopping device includes a local job input unit 21 for inputting a job to be executed on the host, a local spool file 24 for storing the output of the local job input unit 21, a job class declared for the job. The first definition table in which the control information is stored with respect to the first name, a local job class definition table 22, and a first request for selecting and executing a job within a multiplicity range based on the contents of the local job class definition table 22 Local schedule means 23, a job execution means 25 for executing a job based on a request from the local schedule means 23, and control information for stopping a schedule of a job class related to a resource in the host when the resource in the host fails. Failure detection means 2 as the first setting means to be set in the job class definition table 22 6 ₁ , 26 ₂ , the local host 20 including the resource class correspondence table 27 and the schedule stop request means 28, the global job input means 11 for inputting a job to be executed on another host, and the job class name declared for the job. On the other hand, a job is selected within the range of multiplicity based on the global job definition table 12 and the contents of the global job definition table 12 as a second definition table in which the control information is stored, and the corresponding local host (local host 20) is selected. A global host 10 including a global schedule unit 13 as a second schedule unit for requesting execution, and a shared spool file 30 in which the output of the global job input unit 11 is stored and the stored job is executed by the job execution unit 25. Communication between the hosts that notifies the job execution unit 25 of the request of the global job schedule unit 13 Means 40.

Here, the feature of the present invention is that the schedule stop requesting means 28 includes means for notifying the global host 10 of the job class name related to the failed resource via the inter-host communication means 40, and the global host 10 Based on the output information of the notifying unit, the control information for stopping the schedule for the notification source host (local host 20) of the corresponding job class is stored in the global job class definition table.
The second setting means to be set to 12 includes a global job schedule stopping means 14.

Here, a plurality of local hosts may exist, and the global host may also serve as the local host.

The operation of the job schedule stopping device having such a configuration will be described.

Table 1 is a global job class definition table of the job schedule stopping device according to the present invention.
, A multiplicity m for each host constituting the multi-host system, and a schedule stop flag for each host.

Table 2 is a local job class definition table of the job schedule stopping device of the present invention.
It comprises a multiplicity n and a schedule stop flag corresponding to each resource in the local host.

Table 3 is a resource class correspondence table of the job schedule stopping device of the present invention, and is a correspondence table between the job class name C and the resources A, B,... Used by the jobs of the job class.

FIG. 2 is a flowchart showing the operation of the job schedule stopping device of the present invention.

In FIG. 1 and FIG. 2, the schedule of the global job is performed as follows. In the global host 10, the global job input by the global job input means 11 is registered in the shared spool file 30 (S1). In this global job,
A global job class name representing the attributes of the job,
A local host name is declared indicating on which local host the job should be run.

The global job schedule means 13 refers to the contents of the global job class definition table 12. The global job class definition table 12 shows the number of jobs (multiplicity) that can be executed simultaneously for each local job for each global job class. For example,
The multiplicity of host 1 for global job class C ₁ is
a m _11.

The global job schedule means 13 selects a global job from the shared spool file within this multiplicity range to the job execution means 25 of the local host 20 to execute the job via the inter-host communication means 40, and (S2).

The local job is scheduled as follows. In the local host 20, the local job input by the local job input means 21 is registered in the local spool file 24 (S1
1). For this local job, a local job class name indicating the attribute of the job is declared.

The local job schedule means 23 refers to the local job class definition table 22. This local job class definition table 22 indicates the number of jobs (multiplicity) that can be executed simultaneously for each local job class. For example, the multiplicity of local job class C ₁ is n ₁
It is.

The local job schedule means 23 sends the job information to the local spool file 24 within the range of the multiplicity.
, A local job is selected, and a request is made to execute the job (S12).

The job executing means 25 executes the local job input from the local host 20 and the global job input from the global host (S13). When all the jobs are completed (S14), the operation is completed.

When a failure occurs in a resource in the local host, the job schedule is stopped in the following procedure.

The failure of the resource is detected by the failure detecting means 26 (S15). Upon detecting a failure, the failure detection unit 26 notifies the schedule stop requesting unit 28 of the resource name (S1).
6).

The schedule stop request means 28 is provided in the resource class correspondence table 27.
To determine the local job class of the job using the resource, and turn on the schedule stop flag corresponding to the failed resource in the corresponding local job class entry in the local job class definition table (S17). ).

The local job scheduling means does not request execution of a local job belonging to the local job class for a local job class for which at least one schedule stop flag is “ON”. This stops the local job schedule related to the failure (S1
8).

At the same time, the schedule stop request unit 28 notifies the global job schedule stop unit 14 in the global host 10 of the job class name via the inter-host communication unit 30 (S19). The global job schedule stopping means 14 sets the schedule stop flag corresponding to the notified local host 20 of the notified job class in the global job class definition table to “ON” (S3, S4).

The global job scheduling means 13 does not request the execution of a global job whose execution is requested of the corresponding local host, belonging to the global job class whose schedule stop flag is “ON”. As a result, the schedule of the global job related to the failure is stopped (S5).

〔The invention's effect〕

As described above, the present invention has an excellent effect that the system definition can be simplified without the need to define the association between the resource failure in the local host and the job class whose schedule should be stopped on the global host.

[Brief description of the drawings]

FIG. 1 is a block diagram of a job schedule stopping device according to an embodiment of the present invention. FIG. 2 is a flowchart showing the operation of the job schedule stopping device of the present invention. 10 global host, 11 global job input means, 12 global job class definition table, 13 global job schedule means, 14 global job schedule stop means, 20 local host, 21
... local job input means, 22 ... local job class definition table, 23 ... local job schedule means,
24 local spool file 25 job execution means 26 ₁ and 26 ₂ failure detection means 27 resource class correspondence table 28 schedule stop request means 30 host communication means 40 ...... Shared spool file.

Claims (1)

(57) [Claims] An apparatus for stopping a job schedule in a global host of a multi-host system comprising one global host and a plurality of local hosts, wherein the local host is provided for a plurality of jobs executed on its own host. A first definition table storing a plurality of declared job class names and a plurality of resources to be used for those jobs, and a first scheduling means for sequentially requesting execution of jobs based on the first definition table And first setting means for setting information for stopping the use of the resource in the event of a resource failure in the local host in the first definition table, wherein the global host is configured to be executed by a plurality of local hosts. A second class that stores multiple job class names declared for the job and the local host names corresponding to those job class names. A job schedule stopping device in a global host of a multi-host system, comprising a master table and a second schedule means for requesting a corresponding local host to execute a job based on the second definition table; Setting means for notifying the global host of a job class name of a job using the resource when a resource in the own host fails, the global host based on the job class name from the notifying means. A second setting unit for setting information for stopping a job execution request to the corresponding notification source local host in the second definition table.