DSXchange

Hi,

I have a particular px job. I have set the multiple instance property for the job. I have 2 seperate shell scripts invoking the same job with different parameters. I want to run the jobs in parallel through these 2 shell scripts but i see that one of the jobs fails where as the other one runs fine.
I call the datastage job in the sehll script as follows.

using dsjob function

which takes the parameters as follows

dsjob -run -mode NORAML <parameter> -warn <warning limit> -jobstatus <project name> <job name>

Is there any way I can set any of the parameters so that I can run the jobs in parallel.

regards,
varsha

Hi Varsha

While calling multiple instance enabled jobs ( or sequences) through shell scripts, you have to mention instance name after the job name as shown in post above.
One more imp thing, for a multi instance enabled job seqeunces, all jobs called in that sequence should be multiple instance enabled and in the job activities, proper instance name should specified.
Manier times in our projects, we are setting these instances to job parameters so that each time a unique instances created as per the change in job parameter.

Send a mail if you need anything else in this regard as we are doing it here!!!

Thankz for the help, and for all those who replied, I am able to run the multiple instances of the job.
I have a doubt if my jobs are using lookups and if I run the jobs as multiple instances and in parallel, is there a possiblilty of the lookups failing.

Explain WHY you have this doubt.

Have you run the job? Does it work? Do your Lookup stages have reject capturing or do your jobs have downstream null handling on the looked-up column(s), so you can determine whether the lookups failed?

ray.wurlod wrote:Explain WHY you have this doubt.

Have you run the job? Does it work? Do your Lookup stages have reject capturing or do your jobs have downstream null handling on the looked-up column(s), so you can determine whether the lookups failed?

The job works fine, I am in the process of running the job with live data from production, on an average the lookup stage processss around 169219 reocrds, my worry is that if multiple instances of a job run, and if lookup stage is used, it should not result in the jobs to fail. I am uisng the null handling and the reject monitering in case of lookups. Once I do the performance testing, may be I would be in a better position on this.

The job should not fail due to the same lookup referenced different instances of the same job. And the target file names should be taken care of if the same file is created in both instances.

Regards
Saravanan