Auto start job subsystems

The subsystems used for auto start jobs are managed in one program, 'Subsystem. Open' (MNS050). The jobs in a subsystem can be started and stopped from the program 'Subsystem Job. Open' (MNS051) described below.

Working with subsystems in (MNS050)

Panel B displays the subsystem. ASJ is the only valid subsystem type. Use option 1='Create' to create a new subsystem.

Use related option 11='Job in subsystem' on a subsystem record to see the connected jobs in 'Subsystem Job. Open' (MNS051).

Panel E contains the detailed information about the subsystem. The 'Delay' field specifies the default delay in seconds that a job "sleeps" between iterations. The value is normally 60 seconds.

The 'User' field displays the ID of the user that usually starts the subsystem and runs the jobs. This setting can be overridden by entering a user in 'Subsystem Job. Open' (MNS051). The value is normally M3. The user value can define different configurations per user on (MNS100/L) (retrieved from (MNS102), (MNS104)). If you have different system configurations, then there will be different subsystems with their jobs. If you have the same system configuration there will be one subsystem.

Working with subsystem jobs in (MNS051)

Panel B displays all the jobs in the subsystem and enables you to add new jobs. You can select F14='Crt standard' to automatically create the default set of jobs, or you can select option 3='Copy' to manually create the jobs.

Option 11='Sel records' displays 'Subsystem Job. Select Records' (MNS052), where you can select specific jobs to be included.

F14= 'Crt standard' creates the standard auto start jobs. This function creates all auto start job for one subsystem. If you want to create auto start jobs connected to another subsystem, you must select option 3= 'Copy', in front of the current job.

F19 = 'Start all' starts all jobs in (MNS051). If there are several subsystems in (MNS050), this function starts all jobs for all subsystems.

F20 = 'End all' ends all jobs in (MNS051). If there are several subsystems in (MNS050), this function ends all jobs for all subsystems.

Option 9 = 'Str job' starts the selected job.

Option 10 = 'End job' ends the selected job.

The 'Status' of a job has the following valid values:

Note: Only jobs with status 20 will run when the subsystem is started.

The 'Delay' field shows the time, in seconds, that elapses from when the job has stopped processing units of work until it should start again. The processing stops when there are no more units of work to process. The delay time entered in 'Subsystem Job. Open' (MNS051) does not affect the batch job manager, CMNGJOB. That time is set in the M3 properties file.

When the auto start job is activated by a work file, the delay starts to look for new data when the queue does not contain any more records, that is, when the 'No rec' field is equal to 0.

When the auto start job is activated by a transaction file, the delay starts to look for new data after all transactions are run and updated.

The 'File' field defines the work file that activates the auto start job. If this record is blank, the auto start job is activated by a transaction file.

The 'No records' field shows the number of records that are queued and waiting to be processed by the auto start job.

The 'Active' field shows if the job is active and in use. Only jobs with status 20 can be active. However, jobs with status 20 can also be inactive for certain reasons.

The 'Select/Omit' (S/O) field shows if a record is selected with option 11='Sel records' to be displayed in 'Subsystem Job. Select Records' (MNS052), where you can restrict the jobs to be included. For some auto start jobs, a subset of units of work can be selected by using 'Subsystem Job. Select Records' (MNS052). This is useful when more than one instance of a job is needed to increase throughput. For example, one job processes units of work for company 001 and another job processes units of work for company 999.

The 'Active change rate' field displays how active the job is during a period of time. It indicates how many status changes occur when the code is executed e.g. READ, READ_LOCK, or CHAIN etc. (same statuses that can be seen in a profiler log). All status changes do not require the same amount of time; for example, SETLL requires almost no time, while a READ_LOCK can take a few seconds. Therefore, the change rate is not an absolute measure of how fast a job is running. It only indicates the frequency of activity in a job.

On (MNS051/E), these fields are displayed:

Related topics