In earlier releases of CICS®, read time-out is ignored for MRO sessions, tasks waiting on an MRO connection can hang indefinitely. If, for example, transactions in an AOR stall or deadlock for any reason the corresponding relay transactions in an MRO-connected TOR are left hanging. Eventually a problem in an AOR can cause the TOR to stall completely. Although you can prevent new transactions from being routed to, or queued for, the offending AOR, there has not been, until now, an easy way to purge every task that is in flight.
Now, CICS enables the READ TIMEOUT value on the transaction profile definition for MRO sessions. The field in the profile already exists. CICS honours it for ISC but used to ignore it for MRO. CICS now checks this RTIMOUT value for MRO sessions. A new abend code is issued if the RTIMOUT value is exceeded for MRO sessions. By enabling the READ TIMEOUT value on the transaction profile definition for MRO sessions, a cause of intersystem queuing (sometimes referred to as "sympathy sickness") is eliminated.
The situations which may benefit from this change include some which, because they result from stalls or deadlocks elsewhere, are unlikely to resolve themselves unaided.
[[ Contents Previous Page | Next Page Index ]]