Administration Guide › How CA Directory Works › The Backbone Concept in DXmanager › How DXmanager Uses the Topology to Configure the Backbone › Sites and Failover

Previous Topic: Sites and Load Sharing

Next Topic: Assigning the Topology Components to the Namespace Partitions - Instantiation
Sites and Failover

Failover is the ability of a router DSA to continue to service queries even when a data DSA becomes unavailable. If the router detects that a DSA has failed, it resends outstanding requests to another DSA that serves the same partition, making the failure invisible to clients.

If no DSAs within the router DSA's site are reachable, the router DSA sends queries to a DSA in another site in the same region.

If a router cannot find a suitable DSA in any site in the same region, then it sends the query to another region. If this happens, the performance is probably significantly lower, but the service is maintained.

Note: Failover is managed by a router, but it requires that the DSAs in the group maintain their synchronicity, which they do by using replication.

More information:

Failover and Failback