Overview of Performing a Primary/Standby DN Switchover

Performing a Primary/Standby DN Switchover

GaussDB supports primary/standby DN switchover in a shard in a distributed instance. You can promote a standby DN to the primary DN in a shard.

Take the distributed instances that are independently deployed as an example. The instance is configured with 3 shards, 3 replicas, and 3 AZs. If you perform a primary/standby switchover on a node in shard 1 and promote the standby node in AZ 3 to primary, the node in AZ 2 becomes the standby node and the node in AZ 3 becomes the primary node after the switchover.

Figure 1 Performing a primary/standby DN switchover

Changing the Failover Priority of DNs

GaussDB provides failover priority on reliability or availability.

• Reliability applies to scenarios that require high data consistency. In these scenarios, if the primary DN is faulty, services are provided only after log replay is complete to prevent data loss. During this period, connections cannot be established.
• Availability applies to scenarios that require uninterrupted online services. In these scenarios, if the primary DN is faulty and no new primary node is selected within 10 minutes, the standby node is forcibly started as the new primary node to provide services to ensure cluster availability. In this way, the cluster can be recovered as soon as possible at the cost of data loss.