IBM Mainframe Forum

Can someone help me reading AUEP dump read . I want to locate the particular code causing AUEP looping abends.

I found the below link for AICA / ASRA abend read but its not working.

http://nersp.nerdc.ufl.edu/~cicswww/trandump.html#aica

CICS dump reading does not change by the dump code -- you read them all the same.

However, you need to be aware that AUEP does NOT imply -- in any way -- that a transaction loop occurred. From the Messages and Codes manual:

AUEP


Explanation: The task has been abnormally terminated because a return code of UERCPURG has been sent to the User Exit Handler by a User Exit Program. The value of UERCPURG is defined by the macro DFHUEXIT TYPE=EP, ID=xxxxxxxx, where xxxxxxxx is the exit point by which the exit program is enabled. This code does not apply to exit points in domains. The exit program returns this value when it has made a request for CICS services using the exit programming Interface (XPI) and when the XPI call has had a RESPONSE code of PURGED. Exit programs must not set UERCPURG return code under any other circumstance.

While the purge may have been due to transaction looping, there is no guarantee that this is your problem.

Thanks Robert. It should be code problem, as when we bypassed that particular record after which the transaction processed all the backlogs without any issues.

I tried to replicate using the bad record in test however I ended up in vain to locate the problematic code as production keep changing and pretty tough.

The program is big and hard to go spot the bad code and hence I thought if there's a way to locate using offset or some PSW's.

Tough times, looping issues are CPU consuming and highly critical.

Oh, the problem is definitely with the code. The question is what is causing the issue in the code. The XPI (user exit program interface) documentation from IBM indicates that the transaction may have been purged OR THE INTERVAL EXPIRED -- emphasis added -- and suggests to review the reason code for more information. Debugging this type of problem can be difficult as the issue frequently occurs in a totally different program than the dumped program, and the normal debugging methods are little help. Your best bet to resolve the issue may be to work with your site support group to capture a full region dump (not just a transaction dump) when the problem occurs again, and have them help you debug the problem.

Thanks Robert. I tried setting up same test data and debugged the program. Took a lot of time to trace but finally able to spot the bug which caused the looping which is extremely rare scenario that was caught. I am making changes to address the issue now. However wish a easy tracing parameter for AUEP abends wud make things better.

Thanks again for jumping and suggesting the ideas, am glad.