IBM Mainframe Forum

[jairajman]

I have set of jobs running in the morning and evening. These set of jobs process 2 input files. The morning process waits for file 1 and if it arrives it use the lastet generation of file 1 and old verison of file 2 and runs. The evening process waits for file 1 and 2 . What I want is if the file 1 does not arrive by 2:00 PM, I want the jobs that run as part of morning process to be cancelled automatically in CA7.

Please provide your inputs on the same.

[dick scherrer]

Hello and welcome to the forum,

Suggest you change the schedule to submit the job when the dataset arrives rather than start the job and have it "hanging around" waiting. This would not ba acceptable on well-namaged systems as it is a complete waste of resources.

Talk with your scheduling people to have the job submitted when the dataset arrives.

[prakash.subramaniam]

Hi,
We cann't automatically cancel a job in CA7. There are two options, one is cancelling a job in TSO SDSF SPOOL area and another one is cancelling a job in CA7 by giving top line command.

1. Cancelling job in TSO SDSF SPOOL area is putting P or C behind the job.
2. Cancelling job in CA7 is CANCEL,JOB=<JOBNAME>,SCHID=XXX,REASON=AS PER THE REQUEST.

Thanks and regards,
Prakash Subramaniam.
9840187992.

[dick scherrer]

Hello,

The multiple, duplicate replies have been deleted.

Keep in mind that it is far better to schedule jobs properly rather than waste system resources having the job sitting unnecessarily in an initiator. What if 10 other processes had a similar design? 100? Very poor design. . .

Canceling jobs should NOT be part of "normal" operations. Well managed computer operations must formally document every time a production job is canceled. . .

[prakash.subramaniam]

Hello dick,
What u r trying to say now?

Regards,
Prakash subramaniam

[dick scherrer]

Hello,

What i'm saying has been consistent all along. . .

Jobs should be managed such that they run when they are needed. They should not be designed to be canceled when it is determined they are not needed.

A plan to regularly cancel running production jobs is a very sloppy plan. It wastes resources and will eventually cause confusion and operational mistakes. This sort of plan is only introduced by people with no professional experience at managing a multi-user data center.

[prakash.subramaniam]

Hi Dick,
Such a way that all the scheduled production job not run successfully. Due to some System inconsistancy and due to worg data there might me chance to get job abended. During that time we have to run the job again. if not needed, we have to cancel the job.

Thanks and regards,
Prakash subramaniam.

[dick scherrer]

Hello,

Due to some System inconsistancy

1. fix any inconsistancy. . .

due to worg data there might me chance to get job abended.

2. validate the data to prevent the abend. . .

[quote]During that time we have to run the job again. if not needed, we have to cancel the job.
[quote]
3. only submit the job if it is needed - do not submit and then determine if the job is needed.

Overall a most poor implementation strategy.

If this strategy is used repeatedly, it will probably cause more problems that the staff will be able to support. . .

[Debabrata Pruseth]

jairajman,

You can do i thing . Make the job time trigeered with specific values in DOTM (due out time ) , LDTM (lead time ) and SBTM (Submit Time). For eample a DOTM of 1400 and a LDTM of 0030 implies that CA7 has to start the job by 1330 so that it can be done by 1400. Ideally SBTM+LDTM=DOTM , so if you are not giving any DOTM also you can give a SBTM = 1330 and LDTM = 0030 , for the job to end by 1400.

In this case if the file1 is not received by 1400 the job times out and abends. As someone will be monitoring the batch daily , in the evening he will cancel the job from CA7 before the next job starts.

I am not a CA7 scheduler so i can't tell how it is done in CA7 , but the CA7 guys in your project must be knowing this stuff.