Handling Structure Failure during System-Managed Processes

How the system handles structure failure during a system-managed process depends on when in the process the failure occurs and which of the structures failed.

Failure of the Old Structure
If the system has not yet committed to the new structure (prior to the Cleanup phase), or, for duplexing rebuild, prior to the Duplexed Established phase, failure of the old structure is reported to active connectors with the Structure Available event followed by the STRFAILURE event. Users should respond to this event by disconnecting from the structure with REASON=FAILURE.

For both rebuild and duplexing rebuild, once the system has committed to the new structure, failure of the old structure is irrelevant. No STRFAILURE events are reported to any active connectors.
Failure of the New Structure
For rebuild, once the system has committed to the new structure (Cleanup phase), failure of the new structure is reported to active connectors with the Structure Available event followed by the STRFAILURE event. Users should respond to this event by disconnecting from the structure with REASON=FAILURE. If the system has not yet committed to the new structure (prior to the Cleanup phase), active connectors receive the Structure Available event to indicate that the old structure can be used.

Similarly, for duplexing rebuild, once the system has committed to the new structure (Switch phase), structure failure is reported to connectors in the same manner.
Structure Failure During Duplex Established Phase
To handle structure failures in a manner that is transparent to the connector, XES temporarily quiesces access to the structure and captures requests that are already in progress. Once duplexing rebuild has fallen out of duplexing back to simplex mode on the unaffected structure, the system can redrive these captured or quiesced requests. Connectors are made aware of the transition to simplex mode through the Structure State Change Notification event.
Structure Failure During Switch or Stop of Duplexing Rebuild
When switching to the secondary (new) structure:
- If the primary structure fails, no structure failure event is presented. The system drives deferred requests to the secondary structure (now a simplex structure) after the switch completes.
- If the secondary structure fails, a structure failure event is presented after the switch to secondary is complete. The system will drive deferred requests to the secondary structure after the switch completes; the requests will be unsuccessful and will indicate structure failure.
When stopping to the primary (old) structure:
- If the primary structure fails, a structure failure event is presented after the stop to primary is complete. The system will drive deferred requests to the primary structure after the stop completes; the requests will be unsuccessful and will indicate structure failure.
- If the secondary structure fails, no structure failure event is presented. The system drives deferred requests to the primary structure (now a simplex structure) after the switch completes.