If a DB2® or NFS failure occur, Blueworx Voice Response client machines handle an SSI cluster in the same way as they handle an SSI cluster that is running HACMP.
This section relates only to the Cisco ICM custom server, not to the application that might be accessing this custom server (the application developer is responsible for this.) The developer is also responsible for the handling of error conditions that occur in the application that are related to DB2 and NFS failures.
If either a DB2 or an NFS failure occurs where an Blueworx Voice Response for AIX client is operating in an SSI cluster, the custom server continues to function normally if the executable is not paged out. In this condition, the operation is unpredictable and the custom server might fail. If it fails, the custom server must be restarted.
When the Cisco ICM custom server informs the ICM of any change in VRU status it does not refer to the Blueworx Voice Response run status. Therefore, it is best that for Blueworx Voice Response clients that are running in an SSI environment that has the Cisco component running, the ‘System Response during Server outage’ configuration is set to ‘Busy-out all telephony channels’. If an outage occurs, all channels are quiesced, and the Cisco ICM custom server reports the condition to the ICM. The ICM, therefore, no longer routes calls to that VRU if the ICM is configured so that it does not route calls to that VRU (for example, all calls in this VRU might come from external callers), this configuration can be set to the user's choices. Such a use for the configuration must be thought about when the network configuration is being designed.