How does HPR switch paths?

When a link or node outage occurs in the path of the RTP connection, the RTP endpoint attempts to find an alternate route to maintain the connection. If an acceptable route is not immediately found, RTP will continue to try to find one until the path switch timer expires. (The path switch timer duration is modifiable. See the HPRPST start option in z/OS Communications Server: SNA Resource Definition Reference for more details on the path switch timer.) While the path switch attempt is in progress, data transmission is suspended (because of the outage in the original path) until the path switch completes.

Additionally, a path switch attempt can be forced by the operator by invoking the MODIFY RTP command. This will force the specified RTP to make one attempt at finding a better route (as specified by the requirements of the Class of Service) for the RTP connection. If no better route is available, the original route for the RTP connection will be unchanged.

The PSRETRY start option can be used to force VTAM® to perform the following tasks:

Periodically attempt to find a better route for existing HPR connections
Immediately attempt to find a better route when a link on this host is activated or changes status

Like MODIFY RTP processing, if a better route is not available, the original route for the connection is unchanged. For both timer-driven and operator-initiated path switch processing, a path switch can be performed to what appears to be an identical path if the path uses a VR-based transmission group. The reason is that VTAM will perform virtual route selection whenever the first hop in the path is a VR-based transmission group, and this underlying virtual route might change from one usage of the VR-based transmission group to the next. VTAM will not verify that a different underlying virtual route is selected, but instead will just perform the path switch in case a change occurred. For more information, see z/OS Communications Server: SNA Resource Definition Reference.

The HPRPSMSG start option can be useful for large installations that have hundreds or thousands of RTP endpoints on z/OS® Communications Server. This start option controls the HPR path switch message reduction function. In the event of a large network outage, this function can prevent VTAM from flooding the system console log with HPR path switch messages. Although this function reduces the amount of path switch information issued to the console log, the summary information attempts to supply the operator with enough information to understand the size and scope of the outage. For more information about the HPRPSMSG start option, see z/OS Communications Server: SNA Resource Definition Reference.