RE: Why is joblog no complete?

To all that have given input on my question/issue - Thank you I appreciate
it. 
I have been out the last couple of days and am now reading through all the
responses and will see if I can get this figured out.


I did do a quick search through all members in the source file and I did
find an execution of ENDJOB from within an RPG program, so I am going to
start there maybe that is being executed with a LOGLMT set to low and used
to end this job prematurely.

Thanks again

John


-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Andy Nolen-Parkhouse
Sent: Saturday, August 16, 2003 2:54 PM
To: 'Midrange Systems Technical Discussion'
Subject: RE: Why is joblog no complete?

John,

This is a mystery.  If a batch job ends with LR on, then the concept of
LOGLMT does not apply.  This parameter really only pertains to the ENDJOB
and ENDPJ commands.  Considering that your messages are being abruptly
truncated, it would seem that the following condition in the ENDJOB
command pertains:

"If the value specified is less than the number of messages already
written to the spooled file, a message indicating that the limit has been
reached is immediately put in the spooled file as the last entry. The
remaining messages on the queue are ignored."  
(From the CL reference for the ENDJOB command).

So unless I'm missing something, an ENDJOB is either being executed by
someone on the system or it is being triggered by a system limit of some
kind.  For instance, if you specified a specific CPUTIME or MAXTMPSTG on
the job's class, then the job would end abruptly when that limit was
reached.

The default for the LOGLMT parameter on the ENDJOB command is *SAME,
which basically means that the command should not change the
characteristics of job ending from *NOMAX.  This is different from those
prompted commands where the system inserts the word *SAME to indicate
that the value should not be changed.  You could, as an experiment,
create a job that logs CL commands and does some stuff in a loop.  Then
kill the thing with an ENDJOB command and see what you get.  It should
default to *NOMAX and I don't believe that there is a way to change this
default behavior.

Two questions:
- What happened in the three and a half minutes between the time the job
entered the subsystem and that message appeared?  What does SFPGMA do?

- What routing entry is referenced for the 'QCMDB' routing data in the
QSYSWRK subsystem?  Does the class referenced in that routing entry for
the QSYSWRK subsystem have any limitations?

Good luck,
Andy Nolen-Parkhouse

> When I prompt on the ENDJOB command I get the following:
> 
> Job name . . . . . . . . . . . .                 Name
>   User . . . . . . . . . . . . .                 Name
>   Number . . . . . . . . . . . .                 000000-999999
> How to end . . . . . . . . . . .   *CNTRLD       *CNTRLD, *IMMED
> Delay time, if *CNTRLD . . . . .   30            Seconds
> Delete spooled files . . . . . .   *NO           *NO, *YES
> Maximum log entries  . . . . . .   *SAME         Number, *SAME, *NOMAX
> Additional interactive jobs  . .   *NONE         *NONE, *GRPJOB, *ALL
> 
> How does one go about finding out what the actual parameter value is
> when it
> shows *SAME when prompted? (I was signed on as QSECOFR when I prompted
> on
> the ENDJOB command.)
> 
> I have looked at other batch jobs and I see much longer joblogs
> 
> When a batch end with LR on and Return what determines the LOGLMT?
> I am not using ENDJOB to end the batch job




_______________________________________________
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.