Data Guard forced failover Tips Oracle Database Tips by Donald BurlesonDecember 9, 2015

Oracle Data Guard - Forced Failover

If archived redo logs and online redo logs are damaged in the primary database failure, a certain amount of data loss is unavoidable. The forced failover operation is achieved using the ALTER DATABASE ACTIVATE Oracle instance statement. If forced failover is used, the other standby databases in the configuration should be re-created from the new primary database. The forced failover option in the failover to a logical Oracle instance section will be revisited later in this chapter.

            Always perform a graceful failover when failing over to a physical Oracle instance irrespective of the data protection mode of the original primary database.

Oracle instance for Role Transition

One of the most important tasks in the failover scenario is to find the best suitable Oracle instance in the Data Guard configuration to be used for failover. The rules governing the best suitable Oracle instance largely depend on the business requirements and the trade off between loss of data and the amount of time allowed for failover. In general, the following guidelines can be used to determine a suitable standby database:

• Rule 1: A physical Oracle instance should be always given precedence over a logical standby database. This usually minimizes the loss of data and the impact on other standby databases in a Data Guard configuration.
   

• Rule 2: A physical Oracle instance with the highest level of data protection should have higher preference. This will ensure minimal loss of data.
   

• Rule 3: Choose the physical standby database that the most recent archived redo logs have been applied. This will ensure minimal time for recovery.
   

• Rule 4: A logical Oracle instance should be chosen only if there is no physical Oracle instance in a Data Guard configuration or a physical Oracle instance cannot be used for some other reason. Rule#2 and #3 remains the same for a logical standby database.