Run the job using Internal Reader

sanupadh · New User Joined: 29 Jun 2010 Posts: 11 Location: chennai

Hi,

I have job 'A' which runs daily but sometimes abended.I need to create a job 'B' which run after the job 'A' and check if job 'A' get abended then restart it from a 5th step (step05) and also override some parameter of
job 'A'.
also repeate the process untill job 'A' run successfully.
:P

vasanthz · Posted: Thu Jun 16, 2011 2:52 pm

What if the job A never runs successfully?

or you accidentally make a mistake in JOB A step 5, like coding a invalid DSN name and go home for the day, then the internal reader would be like inception movie on ur mainframe, with several levels inside levels inside levels..

Nic Clouston · Posted: Thu Jun 16, 2011 2:55 pm

What is wrong with your scheduling system?

Bill Woodger · Posted: Thu Jun 16, 2011 3:02 pm

Can you please state, in as much detail as possible, and including the JCL for A in "code tags" (grey button that says Code), what your requirement is, rather than what you think is the solution to your requirement.

Anuj Dhawan · Posted: Thu Jun 16, 2011 6:15 pm

Are you talking aout a real, a production, a practical requiremnet?

dick scherrer · Posted: Thu Jun 16, 2011 8:14 pm

Hello and welcome to the forum,

For someone to be able to help, you need to answer the questions asked and provide the requested JCL.

sanupadh · New User Joined: 29 Jun 2010 Posts: 11 Location: chennai

job 'A' run successfully but some times it abended due to specific abend
('U0617' IN STEP 'step5') we always have to monitor it manually and check if it is abended due to 'U0617' IN STEP 'step5'. then need to restart the job from step5 with GDG (0).
We do it manually everyday.
Now my requirement is to create a monitoring job 'B' that run after 'A'
and validates its spool file and do the above work automatically.
for this, below portion I have done.

1.copied spool file of job 'A' in PS file(let it A.spool) using SARBCH utility.
2.I am checking the file A.spool for the return code of job 'A' if it is '0' then send notification mail that job 'A' run successfully.
3.if not then checking if it is abend 'U0617' IN STEP 'step5' then restart the job 'A' using Internal reader.
below is the code

//STEP08 EXEC PGM=IEBGENER
//SYSPRINT DD SYSOUT=*
//SYSUT1 DD DSN=ojb.library(A),DISP=SHR
//SYSUT2 DD SYSOUT=(,INTRDR)
//SYSIN DD DUMMY

now problem is it start job from step 1 but I want to start it form step5
and also I want to change the GDG(+1) to GDG(0).
GDG(+1) using in job 'A'.

gylbharat · Posted: Fri Jun 17, 2011 11:34 am

How would the second job 'B' kick off if the job A has abended?

sanupadh · New User Joined: 29 Jun 2010 Posts: 11 Location: chennai

we will sheduled the job B to run after A everyday.

gylbharat · Posted: Fri Jun 17, 2011 11:57 am

sanupadh · New User Joined: 29 Jun 2010 Posts: 11 Location: chennai

noproblem with this if it is completed or abended (fails) in both the cases
spool file will be created we have to check the spool file for jon 'A'
in case of abend job 'A' terminated with abend.

gylbharat · Posted: Fri Jun 17, 2011 12:15 pm

No that is fine... But Job B will have a dependency in the scheduler.

sanupadh · New User Joined: 29 Jun 2010 Posts: 11 Location: chennai

I dont know much about scheduler but we can schedule job 'B' to run after 'A' wheather 'A' completed or terminated.
or if it is not possible then we can manually run it after 'A'
in case of manual run it can also save our time.
because checking job 'A' manually we have to follow below steps.
check job 'A' in spool .
send mail if successfull.
restart it in case of abend,again send mail .
doing above we can directly run job 'B' it will handle it.
so my problem is not that how can I schedule it but how to create it.

thanks

enrico-sorichetti · Posted: Fri Jun 17, 2011 12:52 pm

why not use Your scheduler exception handling facilities ?

schedule job A
if everything is ok schedule the next normal job
if abend have the scheduler schedule the recovery job

if the recovery jobs succeeds schedule the next normal job
in case of error, then it is time for a manual intervention!

with a scheduler it is a bad practice to restart a job, even more in this case
where You have to skip some steps and change a dataset parameter,
a jcl restart needs a jcl change

it is better to plan in advance,
a proper scheduler setup will handle everything in a much smoother way!
for critical situations it is a normal practice to define/prepaer the related recovery jcls in order to avoid manual modifications to the original jcl

Anuj Dhawan · Posted: Fri Jun 17, 2011 1:57 pm

expat · Posted: Fri Jun 17, 2011 2:13 pm

Anuj, shame on you for suggesting the logical approach

Anuj Dhawan · Posted: Fri Jun 17, 2011 2:24 pm

LOL - actually, I had edited one more sentence for my previous reply, kinda awkward, however, I'll post it now:

PS. Is not it something like, I don't like it when my wife is angry so I look for another girl for that time; when She is normal, I get back to her...

.

No offences meant, just an effect of another book I was reading these days...

sanupadh · New User Joined: 29 Jun 2010 Posts: 11 Location: chennai

Because it abends due to CICS.sometimes problem may happen in CICS region.we cant take care of that except resubmit the job

sanupadh · New User Joined: 29 Jun 2010 Posts: 11 Location: chennai

Abend is due to CICS region which can be down sometimes so we cant take care of that in the JOB 'A' itself.

I heard about IEBEDIT using it we can copy(submit) job to internal reader from perticular step.
but the again problem is how can I override GDG.

Anuj Dhawan · Posted: Fri Jun 17, 2011 2:35 pm

Suggest you tell us a little more about the situation at your end. Creating another Job which will look at the log file from abened or not-abended Job, does not really sound like a plan - just a patch to fix, which might blow at any time.

What would you do, if by some reason, your Job B also got failed someday?

Anuj Dhawan · Posted: Fri Jun 17, 2011 2:48 pm

Looks like you've edited your post about IEBEDIT.

Well, if you insist have a Job B:

As you said, if the job abends or not GDG is always calaloged anyways, when it reachs to you for a solution. Do the following:

- first step will take the GDG(0) to copy to a QSAM-file(flat file).
- copy step from your Job A as-it-is and make use of this QSAM file instead of GDG.
- In Job A have a step to submit above two steps using INTRDR (via Job B) only when there is an abend.

sanupadh · New User Joined: 29 Jun 2010 Posts: 11 Location: chennai

job B is nothing but a monitoring tool.
Below is my JCL.

sanupadh · New User Joined: 29 Jun 2010 Posts: 11 Location: chennai

problem is in STEP07 where now I am running a dummy jcl.
if abend occurs.
but I need to replace the dummy job by job ZFV1710 from step C5171035 with GDG(0)
how can I do this?

sanupadh · New User Joined: 29 Jun 2010 Posts: 11 Location: chennai

Please dont confuse with return code 08.Its OK

Anuj Dhawan · Posted: Fri Jun 17, 2011 3:07 pm

Show us, what is there in: FVTEST.W.ADCXZNK.CARDLIB(DUMMY)?