IBM Mainframe Forum

[GAH]

Little background info:

We have a customer who has 2 jobs that are dataset triggered. The second job has a successor requirement of the first job. The problem lies in the fact that these jobs write to the same named dataset and send it to the same target location, and run within minutes of each other.

My question is, is there a way to set a time delay on the job so once it hits the CA7 queue it waits a desired time to run?

I have reviewed the CA7 Primer and DB maintenance guide with nothing seen to possibly have this accomplished.

Thanks
GAH

[Robert Sample]

We have a customer who has 2 jobs that are dataset triggered. The second job has a successor requirement of the first job.

I'm not sure this makes sense. Why not have the one job dataset triggered, and it triggers the second job? Would that not accomplish what you want?

[GAH]

No because it would still run to close together. Im not sure why these are data set triggered and 1 is a successor. Successor meaning it must successfully complete in order for the second job to start in CA7. Doesnt necessarily mean it has to be triggered by the job nor even be in the same sequence of jobs in the string.

These jobs ran into a problem where a FTP file was much smaller than usual and it caused the system to believe it was writing the exact same information to a dataset due to the fact they run the same name to the same area.

This was a fluke occurence apparently and the user for now is satisfied with leaving it as is.

If there is an answer to a time delay for triggered jobs Id still be interested in hearing it.

Thanks
GAH

[Robert Sample]

I am by no means an expert on CA-7. I know how to trigger jobs by time of day, predecessor job completing, data set creation -- but I don't recall anything able to set how much time between jobs executing. AFAIK, CA-7 considers a successor job as ready to run as soon as the predessor job completes, unless you put a time-of-day requirement on the job.

And terminlogy is critical to IT, where similar terms may mean very different things. The Glossary in the CA-7 Job Management Primer manual specifically provides the definition for "successor" as

successor. A job that runs after another job or an
output network whose tasks must be performed after
a job ends on the CPU.

Your definition of

Successor meaning it must successfully complete in order for the second job to start in CA7.

is not the same -- and may cause you problems eventually.

[oldeagle]

Your best bet is to use CA-7s Automated Restart Facility (ARF). Define a job completion (JC) ARFSET that will take effect if the job completes normally. In the responses section code two actions, first an AW (wait) for the desired amount of time then second, an AC to issue a DEMAND command for the second job in the string.