Rexx to execute spool commands

upendrasri · Active User Joined: 28 Sep 2017 Posts: 121 Location: India

Hello All,

My apologies for the above 3 posts for same question, Some problem with my internet web page after submitting first 2 messages it showed page not found even I didn't get them after login to this forum.

Hope you will understand.

Thanks!

upendrasri · Active User Joined: 28 Sep 2017 Posts: 121 Location: India

Hi Steve and Willy Jensen,

I will explain why I need exact PCT Value.

Some time back we had a spool shortage issue in our environment, After our verification we came to know that thousands of user jobs submitted by the app users. Though each Job is using very less spool percentage but because of the volume of the Jobs were high so we had that issue.

Now am writing a rexx which will lists user jobs and their spool usage , We just keep an eye on those Jobs on daily basis and in case if we feel number of Jobs are more , we will share this report with users and ask them to take copy of their Jobs and delete them from spool.

We can achieve this by using Jes commands against each Job but number of jobs are more and we cannot hide commands/outputs from spool.

Thanks!

steve-myers · Posted: Sun Oct 01, 2017 8:59 am

upendrasri · Active User Joined: 28 Sep 2017 Posts: 121 Location: India

Hi Steve,

Am totally accepting whatever you said. We have aging rules in our environment but After we had that Spool shortage problem, We have given suggestion to App team if they use a specific job class, all their Job outputs will archive to JMR . But its really tough for us to talk and convince business team about these things until we have some data in hand.

With this we want to monitor for few months and see how App jobs are floated in Spool .

Thanks!

Nic Clouston · Posted: Sun Oct 01, 2017 3:27 pm

If, after DUMMYing all sysout DD statements or writing them to data set, you are still getting messages on the spool then look at the allocations for the step. There may be dynamic allocations and you could try overriding them. I have never tried that so I do not know if it will work.

The other thing to do is to run a second job that purges your first job - or simply manually delete it.

steve-myers · Posted: Sun Oct 01, 2017 8:03 pm

JMR is a useful tool, but quite limited. I was at a shop that used it in the 1990s. Within its limitations the shop was reasonably happy with it. That shop also used an IBM output management product - its name escapes me now - that I concluded at the time was an awful piece of junk. It probably could have used a better product, but it was never researched and I'm not sure it was needed.

In the early 2000s, when I got more acquainted with more serious output management products, I took to thinking that JMR is a poor man's output management product.

In the early 2000s I got a job doing support for a more serious output management product. While doing this I had to think about the fundamentals. JES2 is absolutely superb at intercepting and storing output. For a fact it is far, far better than the product I supported. What it's not so good at is management of the output, which you have already discovered. SDSF and similar products like IOF help, but they are limited by thinking in JES2 terms, rather than a broader understanding of the requirements of output management. Of course, not stated, is there is no real broad consensus about this area. People tend to be limited by the peep hole of thinking about their needs and looking at products that appear to satisfy their needs.

Now a forum like this is not a good place to analyze the problems you have, or to recommend solutions. In fact our charter forbids recommending products. Given your concern about analyzing SPOOL utilization less than .01% it seems to me you are being nibbled to death by output held for TSO. Traditional output management products probably won't help you if that's the case. They are more concerned about production output.

upendrasri · Active User Joined: 28 Sep 2017 Posts: 121 Location: India

Hi Nic,

Thank you . Am not aware of how to override those dynamic allocations but I will check manuals on it.

We can delete the output of the Job but our concern is the commands which are processed by that Job are still be there in spool after deletion of the Job.

Thanks!

upendrasri · Active User Joined: 28 Sep 2017 Posts: 121 Location: India

Hi Steve,

Thank you for sharing your experience. We are still analyzing how to avoid this problem in our environment.

Thanks!

Willy Jensen · Posted: Mon Oct 02, 2017 1:38 pm

upendrasri · Active User Joined: 28 Sep 2017 Posts: 121 Location: India

Hi Willy Jensen,

I mean am submitting the below Job

1.
//SDSF EXEC PGM=SDSF
//ISFOUT DD DUMMY
//CMDOUT DD DSN=TEST.COMMAND.OUTPUT,DISP=SHR
//ISFIN DD *
SET CONSOLE BATCH
SET DELAY 600
/DJQ TEST0001,SPL
PRINT FILE CMDOUT
ULOG
PRINT
PRINT CLOSE
/*
//

Even If I delete this Job or Archived to JMR , The command(/DJQ TEST0001,SPL) which is executed by this Job and its output will be still there in Spool. And am executing Multiple commands like this.

JMR is retrieving every thing from the job.

Thanks!

Willy Jensen · Posted: Mon Oct 02, 2017 2:08 pm

It depends, are you using good old SYSLOG or OPERLOG (system logger)?
SYSLOG is indeed spool space, but can be and most likely is offloaded and thus deleted.

upendrasri · Active User Joined: 28 Sep 2017 Posts: 121 Location: India

Hi Willy Jensen,

We are using SYSLOG and yes it will be offloaded but not frequently. The problem here is if I issue more commands and monitoring people will raise a concern on it.

Thanks!

Willy Jensen · Posted: Mon Oct 02, 2017 6:17 pm

Well, then it seems that you are back to using the REXX API whith the precision it provides. Maybe add a random number 1-9 after the xxx.xx in your report to make it look better?

Nic Clouston · Posted: Mon Oct 02, 2017 6:49 pm

upendrasri · Active User Joined: 28 Sep 2017 Posts: 121 Location: India

Hi Nic,

Yeah, If in case I don't get any proper solution to hide them from spool or any other alternative , my last option to explain & convince monitoring people.

Thanks!

upendrasri · Active User Joined: 28 Sep 2017 Posts: 121 Location: India

Hi All,

I have another issue am trying to execute a Rexx program using Jcl but its failing with the message

IEC020I EROPT IS 'ABE' OR NOT SPECIFIED

See the below full details .

My Rexx - SPLPERC1

steve-myers · Posted: Mon Oct 02, 2017 8:55 pm

This is another issue, unrelated to your other issues.

However -

//SYSTSIN DD SYSOUT=*

The TSO Terminal Monitor program expects to read TSO commands from the data set specified by the DD statement with DD name SYSTSIN. I suspect that is rather difficult here.

Nic Clouston · Posted: Mon Oct 02, 2017 9:09 pm

You have been asked to use the code tags when posting stuff sfrom the screen - code, JCL, data etc. Please do so or you may find your topic locked.

upendrasri · Active User Joined: 28 Sep 2017 Posts: 121 Location: India

Hi Steve,

Yeah this is other issue with My JCL. I have modified my JCS as below

//STEP1 EXEC PGM=IKJEFT01,PARM='%SPLPERC1'
//SYSEXEC DD DSN=IGCF.TSO.EXEC,DISP=SHR
//SYSTSIN DD DSN=TEST.CMD,DISP=SHR
//SYSTSPRT DD SYSOUT=*
//ISFOUT DD DSN=TEST.INPUT1,DISP=SHR
//ISFIN DD DUMMY
//SYSIN DD DUMMY
/*

TEST.CMD File contains the below commands
SP
++ALL

But still getting the same error.

Thanks!

upendrasri · Active User Joined: 28 Sep 2017 Posts: 121 Location: India

Hi Nic,

Sorry, Can you please let me know what do you mean by code tags? Am bit new to this forum.

Do you want me to open a new topic with new name?

Thanks!

steve-myers · Posted: Tue Oct 03, 2017 12:42 am

Use a search engine for phpbb code and follow some of those links. phpbb is the software that runs this forum and many other similar forums on the internet. When a moderator suggests you use code tags they mean to use phpbb tags that cause the software to emit ...

Willy Jensen · Posted: Tue Oct 03, 2017 1:53 pm

You say that the modified job ends with the same error messages. But messages from the first job had:
IEC020I 001-5,TESTSPLJ,STEP1 ,SYSTSIN ,JES
I don't think that you get that message again after you assigned SYSTSIN to a dataset.
SP and ++ALL belongs to the ISFIN DD, so I think that you JCL should look like this:

enrico-sorichetti · Posted: Tue Oct 03, 2017 2:41 pm

topic locked ...
a conclusion was not reached after 46 posts , I do not see any chances of getting one