IBM Mainframe Forum

Hi everyone,

I am fresh out of college and have landed a job in the mainframe world as a systems programmer. So far, there is a lot to learn!

I am trying to automate 4 jobs that use the same PRG. These 4 jobs need to run one after another, on a weekly basis. What it should do it XMIT a Dataset from 3 different groups to 1 central group.

There is one PRG to XMIT group 1 , group 2, and group 3. We currently comment out 2 of the 3, XMIT one group, wait for finish and proceed onto the next group, manually commenting/un-commenting in JCL

We have talked about making a slip trap to recognize a unique message in each job and have that take action and start the next job. I suggested making PRG1, PRG2, and PRG3, then somehow use a slip trap to activate the other jobs.

Maybe add a unique line to each job and use that as a slip trap? How involved is that?


What are your suggestions? Slip traps sounds reasonable to me

Excuse my lack of knowledge, please ask for more information if needed.

Thanks!

Your explanation is clear as mud,

the PRG ( program if the abbreviation means that ) is irrelevant to the question

from a scheduling point of view the LUW ( Logical unit of work ) is a JOB

speak to Your support about automating the process using Your organization scheduler product

but what is the reason to run them in strict sequence ?
from your description the jobs could run in parallel

I would talk to the person in charge of AF/OPER for automation and getting it going. Eventually I will talk to him and get some insight as well.

I am in 'training' and I find that asking questions on forums have helped me develop skills in both PC and automotive technical work I do on the side.

I asked the person I am working with on this, why not just run them all concurrently and he said good questions. It is not possible due to some database they all utilize. It scans all the volumes in the environment, pulls data collected for each product. The average time for these jobs are 5hrs.

Does you site have any other jobs that are run with a scheduler? Like the nightly/daily batch process? Mine does, and they run a development version of it too, for running test batches.

Are the jobs running on different boxes? If not, can you just make a single job with the four steps?

To be honest, this ALMOST sounds like a test you've been given to see how you handle yourself.

dr350 wrote:Hi everyone,

I am fresh out of college and have landed a job in the mainframe world as a systems programmer. So far, there is a lot to learn!

I am trying to automate 4 jobs that use the same PRG. These 4 jobs need to run one after another, on a weekly basis. What it should do it XMIT a Dataset from 3 different groups to 1 central group.

There is one PRG to XMIT group 1 , group 2, and group 3. We currently comment out 2 of the 3, XMIT one group, wait for finish and proceed onto the next group, manually commenting/un-commenting in JCL

We have talked about making a slip trap to recognize a unique message in each job and have that take action and start the next job. I suggested making PRG1, PRG2, and PRG3, then somehow use a slip trap to activate the other jobs.

Maybe add a unique line to each job and use that as a slip trap? How involved is that?


What are your suggestions? Slip traps sounds reasonable to me

Excuse my lack of knowledge, please ask for more information if needed.

Thanks!

You've landed a job as a sysprog and are looking at using a SLIP to run a batch job? seriously?
so, do you understand what SLIP processing is and what it is used for?

you really need to look at using a scheduler or some type of automation, not SLIP processing

I, along with others are in career development program that puts us in different sectors within our infrastructure. I was assigned to the mainframe along with 2 others. The purpose is to train us on their roles as the baby boomers are looking towards retirement. The team and everyone in the infrastructure understand this. Many companies are in the same boat.

I didn't go to school and learn mainframes. I should have said that first. I am learning everything from scratch. We have a mainframe guru sitting right next to me but like I use the net for information as well.

better change chair, your guru seems pretty incompetent not to point You to
http://www.redbooks.ibm.com/abstracts/sg246366.html

and from there You can proceed to
http://www.redbooks.ibm.com/abstracts/sg246981.html

( the above link is the first one of the 13 volumes series )

GOOD READING