Predecessor job posting problem

[Haneef Syed]

CA7 job and its predecessor job both are running multiple times in a day.
We tried to add lead time with n hrs  and also zero hours.
But in both cases, If there are two instances of job in Q , both were running one after the other, even only one predecessor job completed at that time.

My assumption is that , only one job instance will run after completion of predecessor job, but here both instances ran at time.

Please suggest how to resolve this issue.

[Roderick Woods]

Haneef,   

CA 7 is working as designed in the scenario you described.  

- you have JOBA that is dependent on JOBX.
- now, you have 2 instances of JOBA in the request queue and both have a dependency of JOBX
- once JOBX start and completed, the dependency for both instance of JOBA will be satisfied.
 
Once in the queue there is no way to prevent both from being satisfied. If you don't want both instance of JOBA to run, you will need to look at some manual or automated way of preventing both jobs from being in the queue at the same time.

There are many ways to accomplish this, but without know the exact scenario I can't give you a simple solution.

[Haneef Syed]

Thanks for reply.

Here is the scenario.

JOBA is calendar schedule jobs runs @ 13:30,15:30,, 17:30, 19:30, 23:30
and dependent, job JOBX  triggered by dummy DSN and runs almost near to JOBA and runs approx 13:38, 15:38, 17:38, 19:38 & 23:38

As CA7 , scan calendar schedule jobs, we are receiving first 2 jobs instances 13:30,15:30 in Q at same  time and same way for next scan 17:30, 19:30 

when job JOBX runs @ 13:38, both JOBA instances are running @ 13:38 and 15:30  <--- Here 15:30 instance rqmt will be satisfying and runs per submit time.
15:30 instance should run after 15:38

Please suggest some solution.

Thanks.

​​​​​

[Roderick Woods]

Since you state JOBA must run after JOBX, I would remove the schedule for JOBA and have JOBX trigger JOBA.

[Haneef Syed]

I was thinking that  Roderick, but problem is that for each run with schedule id(1-5) , there is separate down stream joblist. Please suggest different approach. Thanks.

[Roderick Woods]

Triggering is the way to go. You will have JOBX run under different schid's and it will trigger JOBA because we know JOBA need to run after JOBX. You will have to address how JOBX will come into the queue (scheduled or triggered)  with different schids. You stated JOBX is currently triggered by a dummy dataset. You may want to look at removing that trigger (being the dataset is a not used by job) and just schedule JOBX with multiple schids.
 

[Roderick Woods]

Another option would be to use different dummy dataset names to trigger JOBX with different TRGIDs.

[Haneef Syed]

Yes, having different dummy DSN, which in turn trigger different job names, will work. I will check this.

We can add these different job names as rqmt to schedule job.So that
JobA with schid # 1, will have rqmt of JOBX1
JobA with schid # 2, will have rqmt of JOBX2
...
..
JobA with schid # 5, will have rqmt of JOBX5

I am checking the way, where we can use some option without disturbing  the current setup.
If this is not possible, I will definitely use above option.

Thank you for your time and giving suitable options.


​​​​