HADR reads on standby feature

You can use the reads on standby capability to run read-only operations on the standby database in your High Availability and Disaster Recovery (HADR) solution. Read operations running on a standby do not affect the standby's main role of replaying logs shipped from the primary database.

The reads on standby feature reduces the total cost of ownership of your HADR setup. This expanded role of the standby database allows you to utilize the standby in new ways, such as running some of the workload that would otherwise be running on your primary database. This, in turn frees up the primary for additional workloads.

Read and write clients continue to connect to the primary database; however read clients can also connect to the read-enabled standby, or active standby, as long as it is not in the local catchup state or the replay-only window. An active standby's main role is still to replay logs shipped from the primary. As a result, the data on the standby should be virtually identical to the data on the primary. In the event of a failover, any user connections to the standby will be terminated while the standby takes over as the new primary database.

All types of read queries, including scrollable and non-scrollable cursors, are supported on the standby. Read capability is supported in all four HADR synchronization modes (SYNC, NEARSYNC, ASYNC, and SUPERASYNC) and in all HADR states except local catchup.