The RCM is notified of an Adaptive Server failure when login requests to the Adaptive Servers fail, or when existing connections to the Adaptive Servers fail. Depending on the type of end user and the Adaptive Server, the RCM performs the following processes:
Active Adaptive Server – if an application end-user connection fails, the OpenSwitch server notifies the RCM. If the active Adaptive Server has failed, the RCM starts the failover process. All application end-user connections are suspended until the failover process is finished.
If a DSS user connection fails, the OpenSwitch notifies the RCM. If the active Adaptive Server fails, the RCM routes the connection to the next available server. If there is no “next” server because the other server in the environment is down, the RCM logs an error message. Because DSS users are read-only, the RCM switches them to the standby server without starting the failover process.
Standby Adaptive Server – if an application end-user connection fails, OpenSwitch notifies the RCM. If the standby Adaptive Server fails, the RCM routes the connection to the next available server. If there is no “next” server because the other server in the environment is down, the RCM logs an error message. In this scenario, application end users are working on the standby server because the active server has already failed. The RCM cannot continue to route users unless the active Adaptive Server is running again and able to take login requests.
The RCM and Replication Server support fail over to
two servers only.
If a DSS-user connection fails, the OpenSwitch notifies the RCM. If the standby Adaptive Server fails, the RCM routes the connection to the next available server. This can include routing the DSS user to the active Adaptive Server. If there is no “next” server because the other server in the environment is down, the RCM logs an error message.
See “Failover processing” for more details about failover.
See “End-user connectivity” for more information about users and user connections.
