by Hi all:
I am having a problem with a CICS region configuration and I am hoping
that someone will be able to lend some assistance. Here is the
situation:
AIX version 4.3.3.0 ML 10
CICS Release Number 4.3.0 Revision Level TxSeries 4.3.0.5
We are a medium sized company in Canada and have primarily IBM
hardware and software.
Physically, we have 1 partition for 7 CICS regions. Since managing
the monfile outputs, we have had no space issues. I have created 2
other different regions and they went well, and continue to work well.
I created them by exporting a region and then importing it to create
the new region, and then adjusting environment and misc.
configurations. We have CICS integration with DB2 and Tuxedo. The
lastest region I created we call cicst2.
When I add a program to the region configuration, I add the
transaction id into the TD, and the program and map into the PD
database area. This has always worked fine.
Here is the problem: the region can not seem to locate the cobol
programs. It errors and shuts down the region when I try to add a
startup program to the configuration (successfull in other regions.)
If I remove the startup program and start the region, from within
CICS, I execute:
CEMT SET PROG(WB99QCLN) NEW
and I get an "Invalid change" error. This happens only for the
PROGRAM, but NOT for the map file. I can 'NEW' a map without error.
This is the case on every program (there are about 50.)
If I add that program (QCLN is just a program to clean up TDQs) to the
startup configuration I get this:
09:32:20 cicst2 : Transient Data queues processed!
09:32:20 cicst2 : Processing temporary storage queues ...
09:32:21 cicst2 : Temporary Storage queues have been processed!
09:32:21 cicst2 : Processing local queues that correspond to
asynchronous processing requests ...
09:32:21 cicst2 : Local queues have been processed
09:32:21 cicst2 : The CICS initialization transaction has ended
09:32:21 cicst2 : Checkpointing started
09:32:21 cicst2 : Running startup program 'WB99QCLN'
09:32:21 cicst2 : Unable to find startup or shutdown program
'WB99QCLN'
09:32:21 cicst2 : CICS is performing region abnormal termination
in process 'cicsas'
09:32:21 cicst2 : Dump to 'SYSA0001.dmp' started.
09:32:21 cicst2 : Dump to 'SYSA0001.dmp' completed.
09:32:21 cicst2 : Shutdown of CICS region 'cicst2' is complete
09:32:21 cicst2 : Checkpointing completed
I trimmed off the left side of this output so that it would display
better. The Unable to find error is ERZ010005E/0530. I've done a
search on this and all it seems to question is whether there is a typo
in the name of the program in the configuration definition.
I have checked and rechecked the userids, groupids, owner groups,
permissions and I can't see any inconsistency between this CICS region
and the other working regions.
As a sidenote we have been getting the following error since last
summer in every region. As it has never prevented any processing and
never caused any issues, we have not gone about fixing it. It bugs me
but I believe it is not related to my current problem:
ERZ025001E/0081 03/07/03 09:32:20 cicst2 : Unsuccessful open of
file: '/.:/cics/sfs/t2sfs'/'dfhcfil', index: 'cfilkey', SFS code -
1997905970 'ENC-sfs-0050: File by this name does not exist.'
ERZ021004E/0042 03/07/03 09:32:20 cicst2 : Unable to open file
'FILEA'
[end of sidenote.]
I have compared and examined the CICS configuration settings through
smitty and via the console log files. I can't understand why I can
NEW a map file but not an executable in this new region. (I assume
that if I can NEW it, that CICS actually finds the map file and
imports a new copy into the runtime database.)
As an overview, it appears that the CICS region simply doesn't have
the proper path to find the program. The programs are there, and the
pathing syntax and conguration syntax is mimiked from a working
region.
Something else the bothered me. I have been unable to find within the
CICS configuration or database files, anything the explicity states a
working or chroot directory for the region. The definition of the
program directory is forced to be relative. Our regions are all in
/var/cics_regions/ and the program definitions within CICS define
programs in the cspbin directory.
Structurally, /var/cics_regions/<region_name>/cspbin is a link to
another partition which contain the programs. from the command line
sitting in the <region_name> directory, I can list the programs just
fine.
I can run CESN and CSSF (signon and signoff) from within the CICS
region, but all others fail with:
ERZ014016E: Transaction 'VCLM', Abend 'A141', at 'BILL'.
Without wanting to drag this one, I should mention that I did notice
an IBM PTF fix for an "Invalid change" error when trying to CEMT I
PROG(prog_name) NEW on TXSeries 4.3.0. As this error only happens
when trying to SET PROG and not INQ PROG, and since our other regions
of CICS are working without issue, I have not persued this angle.
Question: do all the cobol programs have to be recompiled if they will
be running in a new/different region?
This is a puzzler.
Any insight into this problem appreciated.
Thanks for reading.
Bill
Regina, Canada