IBM Mainframe Forum

by **Josh Stan** » Tue Oct 19, 2010 1:09 am

Hello.

I am assigned a project that converts a COBOL batch job to CICS. I know that this is usually a bad idea, but the choice isn't mine. Our users don't want to wait until the next day to get their data.

I have a cics program that updates VSAM files, based on what is in a database for 'process-num' when it is called (process-num is a parameter in FUPLAN-COMMAREA and BARORDER-COMMAREA). The program works fine when it is called to update a plan file and when it is called to update an order file. The problem happens when it is called twice to update both plan and order files.

Batch program FUPLAN contains this code:

Select all

EXEC CICS LINK PROGRAM ('FUBAR1')
TRANSID ('EXCI')
COMMAREA(FUPLAN-COMMAREA)
LENGTH (LENGTH OF FUPLAN-COMMAREA)
RETCODE(EXCI-EXEC-RETURN-CODE)
SYNCONRETURN
END-EXEC.

IF EXEC-RESP NOT = 0
SET ABEND-ERROR TO TRUE
DISPLAY 'EXEC-RESP = 'EXEC-RESP' EXEC-RESP2 = 'EXEC-RESP2
SET ADDRESS OF WS-ERROR-MSG
TO EXEC-MSG-PTR
DISPLAY WS-ERROR-MSG
END-IF

Batch program BARORDER has the same code. The difference between FUPLAN and BARORDER is the data they put in the DB.

The error I get varries somewhat, but the following is an example.

Select all

2010-10-18 11:37 05.74 FUBAR1 STARTING EXECUTION
DATE 2010-10-18PROCESS # 05469 PROCESS STAT( )
2010-10-18 11:37 05.81 FUBAR1 STARTING EXECUTION
DATE 2010-10-18PROCESS # 05470 PROCESS STAT( )
FUBAR1 EIBRESP = 00000022 EIBRESP2 = 00000014 ERROR WRITING FILE PLNFILE1
FUBAR1 PLNFILE1 FETCHED 6 RECS 5 WROTE TO PLNFILE1
FUBAR1 2010-10-18 11:37 06.01 W PLNFILE1 ERROR WRITING PLNFILE1
FUBAR1 ABNORMAL EXECUTION
2010-10-18 11:37 06.03 FUBAR1 FINISHED EXECUTION
DFHAC2236 10/18/2010 11:37:06 LRTEST3 Transaction EXCI abend USER in program PLN
resources will be backed out.
FUBAR1 EIBRESP = 00000014 EIBRESP2 = 00000150 ERROR WRITING FILE ORDFILE1
FUBAR1 ORDFILE1 FETCHED 6 RECS 1 WROTE TO ORDFILE1
FUBAR1 2010-10-18 11:37 06.13 W ORDFILE1 ERROR WRITING FILE ORDFILE1
FUBAR1 ABNORMAL EXECUTION
2010-10-18 11:37 06.16 FUBAR1 FINISHED EXECUTION
DFHAC2236 10/18/2010 11:37:06 LRTEST3 Transaction EXCI abend USER in program AIM
resources will be backed out.

Any pointers to the right direction will be greatly appreciated.

by **Josh Stan** » Tue Oct 19, 2010 3:00 am

Update: I didn't realize that temp storage queues are global. That was my first problem. FUPLAN would kick off an instance of FUBAR1 and write to the tsq then BARORDER would start an instance of FUBAR1 that would write to it. They then read from the queue and since the data got jumbled in the same queue, unpredictable errors occurred.

I now have a length error that seems to only occur when they are run together, but it is more isolated. I will dig into it some more before posting(only posting if I can't find the answer).

If you got this far, thanks for reading.

by **dick scherrer** » Tue Oct 19, 2010 3:18 am

Hello and welcome to the forum,

Thanks for the update - good luck with the rest

Someone will be here if there are "opportunities".

d

IBM Mainframe Forum

link in batch works, 2nd link fails both, parallel problem

link in batch works, 2nd link fails both, parallel problem

Re: link in batch works, 2nd link fails both, parallel problem

Re: link in batch works, 2nd link fails both, parallel problem