CICS: A scenario for a AFCY abend

The below problem will occur when we use LSR buffering for KSDS file in CICS.


Step 1: A successful EXEC CICS STARTBR is issued for a KSDS file

Step 2: Successsive READ NEXT is issued for the same dataset

Step 3: After checking the condition , a decision is made to update or

delete the record read on the READ NEXT operation and an EXEC CICS READ

UPDATE or EXEC CICS DELETE is issued.

The above operation will lockout the task and force it to hang and the

reason is :

The CI's that are accessed by a task or enqueued (locked) in an LSR buffer

pool and dequeued (released) either when the task is finished or an explicit

release request is issued. So in the above example the CI's will be locked

during the READ NEXT operation and the CI's won't be released until the

ENDBR command is issued. So when you try to update or delete the record ,

the record sitting in the CI is not available for the same task again and so

the task will hang till purged or cancelled by CICS which will cause the

ACYA abend.