An Adaptive Server failure within the high availability, warm standby environment occurs if login requests or existing connections to the Adaptive Server fail. If the Adaptive Server fails, the OpenSwitch server passes the notification to the RCM.
The RCM attempts to connect to the Adaptive Server.
If the attempt fails, the RCM logs the server as DOWN.
If the attempt succeeds, the RCM determines if the requested database is available by monitoring database connections.
If the requested database is listed in the REQUIRED_DBS configuration parameter, the RCM attempts to connect to the database. If the attempt fails, the server is considered down. If the attempt succeeds, the server is considered up.
If the requested database is not in the list, the RCM considers only the status of the server and not the database when pinging the Adaptive Server. Because the server status is UP, the RCM does not begin the failover process.
This two-step process gives you finer control over failover. For example, you can prevent noncritical databases that become unavailable from starting the failover process.
Adaptive Server allows users to connect to the server
even if the requested database is unavailable.
End users receive an error message, but are still connected to the
server. This means that the Adaptive Server does not notify the
OpenSwitch server and, therefore, the RCM, when users attempt to connect
to a database that is unavailable. However, the RCM is notified
by the OpenSwitch server when existing connections fail because a database
has become unavailable and the RCM can start the failover process.
