When failover or failback occurs, all existing client connections are lost. As failback is the same as failover, clients do not see any difference between the two events and the connection failover happens the same way in both cases. However, the client connection failover happens differently depending on the type of connection the client has established with Adaptive Server. Client connections are either HA connections or non-HA connections.
HA connections have the CS_HAFAILOVER property set in the connection handle, and the hafailover entry in the interfaces file. For clients that use the HA connection, failover is transparent. That means that the broken connections are automatically re-established when the Adaptive Server restarts on the secondary node. However, the client must resubmit any uncommitted transactions.
Non-HA connections are regular connections and do not reconnect automatically. With non-HA connections, the clients must first re-establish their connections to Adaptive Server, then resubmit uncommitted transactions.
For more information see, “Configuring the interfaces file on the client side”.
