Conference rdgeng::cics_technical

164.0. "Program cache in MF Cobol" by DEKVC::DOOGYUPARK () Wed Jan 29 1997 08:07

Hello,

I tried to enable program caching in MFCobol v3.2b,cics v2.1a,unix v3.2e.

Generally speaking,it works but have a problem like this:

	1. set COBSW=+l70000000-N-O-F+A in env. file.
        2. modify a program.
	3. Make new copy of the program in CEMT.
	4. Start a transaction of the program.
	   At first starting,i always get an error,
	 
	   ERZ014016E/0028 Transaction 'KM00',Abend A158,at LK00,

	   and the transaction abnormally terminated.

	   At next starting again,thers is no error and it works.

After modifying a program and doing new copy,i always get this error only
at the first running of the transaction and once after i get this error,if 
i run the transaction again,i always can run the transaction successfully.

Do you have any idea ??

Thanks a lots,
dg.

------ appendix : symrecs
SYMPTOMS = PIDS/569725110 LVLS/211 PTFS/01 96/01/10 1.1 RIDS/PinCA_StartCobol LINE/-1 MS/015027 MSN/89 SRC/11 PRCS/2097152 AB/A158 TIME/970129165651 KST
SECONDARY SYMPTOMS = PostMortem (Error Path is PinCA_StartCobol<TasPR_CallApplication) logging where error occurred
SYMPTOMS = PIDS/569725110 LVLS/211 PTFS/15 96/01/10 1.1 RIDS/TasPR_CallApplication LINE/527 MS/015027 MSN/89 SRC/11 PRCS/0 AB/A158 TIME/970129165651 KST
SECONDARY SYMPTOMS = Runtime exception (exc_e_illaddr) in TasPR_CallApplication
------- 

DG.

Please can you post the console.msg for around the time you did the CEMT
NEWCOPY.  

There is a SEGV occuring, but the CSMT.out shows it is in the
code that calls the COBOL - unfortunately this means that it could either
be in the CICS code or the COBOL loading code.  The console.msg should 
confirm which area it is happening in.

Thanks.

Gerry

    Gerry,
    
    Nothing is logged in console.msg for around the time i do CEMT NEWCOPY.
    
    And just the response is NORMAL in CEMT NEWCOPY screen.
    
    Thanks,
    dg.
                                                    

    And this symptom looks depend on program by program.
    Some program does not produce this symptom at all.
    
    I tried just do NEWCOPY without re-compiling(modification) a program
    and got the same error.
    And sometimes this program did not produce the error while sometimes
    it produced the error.
    
    Thanks,
    dg.
    

   
Thanks for the information.

IBM Hursley engineering are looking at this failure.  They think that it
has been seen on other platforms.  I'll post another reply just as soon
as I have an update.

-gerry