Optimizing Storage Management › Session Storage Cleanup › Node Error Recovery

Node Error Recovery

You can purge a session by invoking the VPE purge storage function from a Node Error Program (NEP). If available, CICS invokes a NEP when VTAM notifies it that a terminal was lost (LOSTERM notification). The NEP schedules a transaction that issues the purge storage function on behalf of the lost terminal.

To invoke the purge storage function from a NEP, first customize and then add the following statements to your site's CICS Node Error Program:

NEP0AF   DS    OH                                              @BD5021A
*-----------------BEGINNING OF SUGGESTED VPE PURGE INVOCATION CODE----*
        CLI   TWAEC,TCZxxxx       ERROR?
        BE    PURGEIT         
        CLI   TWAEC,TCZxxxx       ERROR?
        BE    PURGEIT
        CLI   TWAEC,TCZxxxx       ERROR?
        B     GO_ON               NO; ELSE...
PURGEIT DS    OH
        EXEC  CICS START TRANSID('VPUR') LENGTH(4) FROM(TWANID)
GO-ON   DS    0H
*-----------------END OF SUGGESTED VPE PURGE INVOCATION CODE----------*

In the preceding illustration, 'xxxx' represents the error code you intend to trap. Commonly trapped error codes include TCZTXCU (node unrecoverable), TCZNSP01 (network error 1) and TCZNSP02 (network error 2). The values that need to be trapped will depend on your network configuration. See the DFHZEQU macro for specific error code definitions. Also see the CICS Customization Guide for more information about Node Error Program processing.