Multinode persistent sessions is used to facilitate session recovery after a node failure. It can also be used to move applications to another VTAM® in the event of an application failure or as part of an MNPS takeover. The application can recover on the same VTAM (if that VTAM is restarted) or any other VTAM that is part of the sysplex and that is connected to the MNPS coupling facility structure. Refer to z/OS Communications Server: SNA Network Implementation Guide for complete information about setting up a sysplex that supports multinode persistent sessions.

For an application to be capable of multinode persistent sessions, PERSIST=MULTI must be coded on the APPL definition statement for this application progam. There are no differences in the programming interface for applications using single-node or multinode persistent sessions, with one exception. For both functions, the application must enable persistence as described in How an application establishes persistence, and for both functions the application can indicate support of forced takeover processing by using the SETLOGON OPTCD=PERSIST macroinstruction. The exception in the interfaces between the two types of persistence is that in order to initiate MNPS forced takeover, an application needs to specify an additional parameter on the ACB macroinstruction, as described in Response to an application failure with MNPS.