IBM Mainframe Forum

Hello all,
I'm out of ideas. LISTDSI function abended with S113-38 (meaning: "An OPEN was issued for a VSAM data set with a DCB, but the format1 DSCB for the data set does not indicate it is a VSAM data set. Ensure that the DSNAME specified on the DD statement is correct and that the program supports opening a VSAM data set with a DCB"). The file type is IAM and the IAM software is up and running, so from the programs' perspective these files should look like normal VSAMs.
The funny part is that I have slightly different version of IAM than the machine where the abend occurs, so I've never seen it live. The Rexx and z/OS versions are the same. Short register dump we've got is useless (no active module found), I have just the dataset name, abend code and the abending function.

Has anybody ever encountered this behavior?
Is there anything I can do, or should we just try different version of IAM and hope it helps?
If nothing else works, is there any way to at least suppress the abend messages? The abend is quite harmless, the exec continues to run and subsequent modules that are being called handle the same dataset without any problems. The big red messages on the screen just don't look good.

Thank you for any suggestions.

did You care to read the REXX manuals for the LISTDSI limitations, looks like not
here is a link to the tso bookshelf
http://publibz.boulder.ibm.com/cgi-bin/ ... s/IKJ4BK90

a listdsi for a VSAM dataset will end with a return code 4 and a reason code 12
which according to aforesaid manual means 12 VSAM data sets are not supported.

You can check Yourself by running a simple rexx like


the best thing to do is to ask Your support and get IAM people involved

not a rexx issue

I agree with enrico-sorichetti that this is not a TSO/Rexx issue, except why is LISTDSI even opening the dataset? If I remember correctly IAM dataset DSCBs look like BDAM, not VSAM. I reviewed the Rexx variables set by LISTDSI, and as far as I can tell the only time it opens the dataset is if it is PDS and the directory option is specified in the argument string. None of the Rexx variables set by LISTDSI appear to be VSAM related.

Yes, I have read the LISTDSI manual:
"If you use LISTDSI to retrieve information about a VSAM data set, LISTDSI stores only the volume serial ID (in variable SYSVOLUME), the device unit (in variable SYSUNIT), and the data set organization (in variable SYSDSORG)."
The "VSAM data sets are not supported" refers to why there would be RC=4 and reason code 12, but nothing about any possible abends.
I'm using the LISTDSI just to determine the DSORG (and maybe some more information if possible) and then decide what to do next.

The DIRECTORY option is on, so the solution may be to add one more test without the DIRECTORY ahead of it and when it detects a VSAM (IAM), skip the next test. Thank you Steve for the tip!

If anyone is interested, LISTDSI without the DIRECTORY parameter abends as well.

Sounds like you need to post the actual part of the code, the output (is it a batch or online job), and the info about the dataset. The error says that your definition and the dataset do not match, maybe you are missing something. How about the amount of storage you are using for the Job/TSO?

LISTDSI works well with standard zOS datasets
this topic deals with IAM datasets

again not a TSO/REXX/LISTSDI issue

The IAM vendor should provide a function similar to LISTDSI.

It is not clear why you would expect IBM products to know about and support specialized dataset orgs from a particular vendor.

Solution: don't use SMSINFO parameter. DIRECTORY has no effect on this.
Case closed.
Thank you all for your help.

Mods, if you think this thread is too offtopic or unhelpful, feel free to move/delete/lock/whatever.