Providing a Response for a Failed Connector

The system permits a connector to respond on behalf of a failed peer connector in two instances:

If the connector failed with an outstanding response to EVENT=REBLDCLEANUP
If the connector failed with an outstanding response to EVENT=REBLDSTOP

and not all event responses have been received.

With the capability to provide a response on behalf of a failed connector, the previous rebuilding scenario could result as follows:

Connection B fails with an outstanding response for a Rebuild Cleanup event.
Connection A is notified of the Disconnected or Failed Connection event.
Connection A responds to the Rebuild Cleanup event for Connection B using the PROXYRESPONSE=YES parameter.
The rebuild process completes.
Connection A responds to the Disconnected or Failed Connection event.

Note that the connector issuing IXLEERSP with the PROXYRESPONSE=YES keyword is responsible for following any protocols that the application uses during its rebuild cleanup or rebuild stop processing. For example, suppose an application uses two structures — a cache structure and a lock structure. When the rebuild of the lock structure enters the rebuild complete phase, the application updates the cache structure. If a connector fails at this point, and a peer connector decides to respond for the failed connector with the PROXYRESPONSE keyword, that connector has to ensure that the updates to the corresponding cache structure are performed. The updates could be performed immediately by the active connector that issued the PROXYRESPONSE confirmation, or could be done during the processing of the DISCFAILCONN event.