Initializing of connections

There is a class of problems that occur because of the order that the MVS™ address spaces become activated during a system IPL. As an MVS system is being IPLed, address spaces reach a state in which they are able to service requests at different rates. Therefore, there is no assurance that the VTAM® and APPC address spaces, that are necessary to service RRSF requests, are available when the RACF® subsystem address space issues its first request.

Attempting to activate the connections or send information to remote nodes when the RACF subsystem address space is started, but before VTAM and APPC have completed initialization, result in APPC failures. Normally, such APPC failures cause the connections between the two nodes to become operative in error when the first attempt is made. Instead, RRSF allows the RACF subsystem address space to continue to poll for APPC/MVS availability and when available, attempts to register the RACF subsystem address space as an APPC/MVS server.

If the RRSF APPC server fails to successfully register with APPC/MVS, error message IRRC023I is issued to the MVS operator console and RRSF initialization discontinues.

A registration attempt to APPC/MVS also occurs if:

A local node is in a dormant state,
A TARGET command is issued to make the connection operative, and
Protocol information has been supplied.

A subsequent TARGET LIST command against the node would show a connection state of operative active.

Note: If MODENAME is not specified in the TARGET command, RACF sends a default name to APPC that allows the connection to complete successfully. This default name, IRRMODE, is placed in IRRNODE. However, since there was no explicit specification by the user and the default name is a RACF internal, the TARGET LIST output shows "NOT SPECIFIED".