IEF450I SECUR02 MATCH01 - ABEND=S04C U0000 REASON=00000000

Adrian Stern · New User Joined: 13 Feb 2008 Posts: 12 Location: Sweden

An extremely simple rexx program I try and run in batch dumps BEFORE STARTING with:
IEF450I SECUR02 MATCH01 - ABEND=S04C U0000 REASON=00000000

SYSTSPRINT says:
IKJ56641I MATCH ENDED DUE TO ERROR+
IKJ56641I USER ABEND CODE 1000 REASON CODE 00000000

The tso program run is IKJEFT1B

The extraordinary thing is when this job step is run on it's own there's no problem but when one of multiple steps (i've tried both first and last) it dumps.

Anyone any ideas?

I even tried executil ts before it but that just got a ready reply!

I've been writing rexx for over 25 years and never seen anything like it

MBabu · Posted: Thu Dec 10, 2009 10:47 am

Are you calling IKJEFT1b directly from an EXEC PGM= in the JCL or is there another program calling IKJEFT1B? There are cases where the some authorization at the job step TCB changes so that one can't invoke additional authorized programs, but unfortunately I don't have the details. This problem just looks familiar to me. I suspect there may be info in the authorized assembler services guide. In any event, what is the program before the IKJEFT1B invocation?

(update: I missed the part about it fails even if it IKJEFT1B is the 1st step. Still, what else is being called?)

PeterHolland · Posted: Thu Dec 10, 2009 12:55 pm

If a command or program being processed by IKJEFT1B ends with a system or user abend, IKJEFT1B causes the job step to terminate with a X'04C' system completion code. IKJEFT1B also returns to the caller the completion code from the command or program in register 15.

Maybe you show the simple rexx code for soem clarification whats happening?

Adrian Stern · New User Joined: 13 Feb 2008 Posts: 12 Location: Sweden

The step is:
//MATCH02 EXEC PGM=IKJEFT1B
//SYSEXEC DD DISP=SHR,DSN=KONS89.REXX.EXEC.BATCH
//SYSTSIN DD *
MATCH 10,1,80
/*
couldn't be simpler.

I could post the rexx code but there's no point as the program doesn't even start! Trace in the rexx shews nothing and executil ts as I sais just gets a ready prompt which is also incredible!

PeterHolland · Posted: Thu Dec 10, 2009 2:20 pm

I guess MATCH is a function, and yes that function call is simple enough.
But not knowing what happens inside that function will make it impossible
to find anything erroneous in that function.

I wish you good luck.

Adrian Stern · New User Joined: 13 Feb 2008 Posts: 12 Location: Sweden

Well I was obviously running in stupid mode! MATCH is such a common name that I should have realised that there was another program with the same name! The complete jcl had a joblib card which was not present when I ran the single step! So renaming my program solved the problem.

Maybe age does imply degeneration?

expat · Posted: Thu Dec 10, 2009 3:32 pm

PeterHolland · Posted: Thu Dec 10, 2009 6:42 pm

Adrian, expat thats so true.
Happens to me all the time, even after i retired (or was it retarded?)

dick scherrer · Posted: Thu Dec 10, 2009 9:23 pm

MBabu · Posted: Thu Dec 10, 2009 10:24 pm

that's why for 25 years I always used a test clist called BLAB... then one day I created a load module of that name and *poof* -- same problem and same "why doesn't this work?"

... happens to us all. But it is a good way to get you into the habit of using the % (eg: %MATCH) to avoid this problem.