Applications using the primary RAPCache do not necessarily failover to the secondary RAPCache, if the primary data stream fails. The failure may not be with the RAPCache itself; upstream components may have failed.
For this reason, the normal Open Client (and ODBC, JDBC) connection failover mechanisms that are available to Sybase client applications cannot be relied upon, unless the primary RAPCache is shut down. Otherwise, when a failure in the primary data stream is detected, software using the primary RAPCache must be redirected to failover by other, non-automatic means.
There are several possible scenarios for minimizing the complications of connection failover:
If the primary RAPCache is running, shut down the RAPCache server. The shutdown forces automatic client-side connection failover, if failover is configured for client applications.
Use Sybase OpenSwitch. OpenSwitch is designed for handling this situation, but must be instructed when to make the failover, as failure of the primary data stream may not be obvious.
Change applications to be aware of both the primary and secondary ASE installations. For existing application software, this may be impractical for the user.
Change the interfaces or sql.ini file of the application. For hundreds of applications running on PCs or workstations, however, this may not be a practical solution.
