Demanding a job into the request queue does not automatically result in immediate submission for execution. When a DEMAND is issued for a job, that job is placed in the request queue where the availability of its input requirements is determined. (To request immediate scheduling without requirements checking, use the RUN and RUNH commands.) The same process occurs when a job is automatically scheduled into the queues.
Using SET=NDB can allow a job not defined to CA Workload Automation SE to be run under the CA Workload Automation SE control and not be added to the database when it completes.
Input workstation networks must be demanded separately from jobs. If a demanded job requires input workstation activity, the network responsible for that activity must be requested separately using the DMDNW command.
Preprocessing work can be demanded well in advance of the job for which it is a requirement. This can be necessary if the preprocessing activity requires several hours or days of lead time, or if the work is available and is requested to be done early during a slack period.
Output workstation networks associated with a job are automatically placed in the postprocess queue when the CPU job is demanded.
Output networks can be demanded separately from their associated CPU jobs by a DMDNW command.
Demanded work is assigned the UID of the operator issuing the DEMAND command if the job is not defined in the database.
If TYPE=RES is specified on the DEMAND command, the QM.4-X panel, XRST, or RESTART commands can be used to restart or rerun the job.
FROMNODE is a reserved keyword that is not considered valid input unless generated by CA Workload Automation SE internally. This keyword cannot be coded on online, batch, or trailer terminal transactions. It is reserved to indicate that the DEMAND is issued in response to a scheduling request from an XPS client.
All internal cross-platform jobs must be defined in the CA Workload Automation SE database because selected information for each job is contained in the internal cross-platform job definition. Attempts to add them using the DEMAND(H) command will fail.
Repeating Jobs
A job can be brought into the request queue using the DEMAND command and executed more than once. Each iteration must successfully complete (or be forced complete) before the next iteration is added to the request queue.
The following keywords are used with repeating jobs:
Required. Defines the amount of time between each iteration of the job.
Required. Specifies how the submit time requirement of each iteration is calculated—by adding the INTERVAL to the previous iteration's submit time requirement (CLOCK), START, or END time.
Optional. Defines the maximum number of times the job will be repeated.
Optional. Defines the clock time after which the job is not to be repeated. If not specified, the stop time is 1439 minutes (24 hours minus 1 minute) after the submit time requirement of the first iteration.
Required. Defines the submit time requirement of the first iteration.
Optional. Defines the amount of time required to process the job. If not specified, 0100 (1 hour) is used.
Optional. Defines the due-out time (expected completion time) of the first iteration. If not specified, the current time plus the LEADTM is used.
As each iteration successfully completes (or is forced complete), CA Workload Automation SE adds the INTERVAL to the submit time requirement, start time, or end time based on the TYPE setting. This value becomes the next iteration's submit time requirement.
The new submit time requirement is then compared to the STOP time. If either the new submit time requirement or the current time are greater than the STOP time, the job is not repeated.
The new job's due-out time is calculated as the new submit time requirement plus the difference between the original due-out time and the original submit time requirement.
The new job's deadline time is calculated as the new due-out time minus the original lead time (LEADTM).
The start and end times of the job are local to the system where the job executed, not local to the system where CA Workload Automation SE is executing. TYPE=START and END should be used with care for NJE and cross-platform jobs.
|
Copyright © 2013 CA Technologies.
All rights reserved.
|
|