BystandersFollowRoleChange

The BystandersFollowRoleChange configuration property establishes whether bystander standby databases are evaluated during failover (value = ALL) or after failover (value = NONE).

Category Description

Broker default

ALL

Configuration Type

Data Guard only

Datatype

String

Property Type

Configurable

Scope

Configuration

Valid Role

Primary, Physical Standby

Valid value

ALL or NONE

Descriptions of the ALL and NONE options are as follows:

  • ALL - During the failover process, the broker determines whether the bystander standby databases are ahead of or behind the failover target standby (that is, the standby that will be the new primary).

    If the bystander standbys are ahead, they will be disabled with a status of ORA-16661 as part of the failover operation. The bystander standbys must be reinstated after failover completes. The broker reinstates a standby through a flashback to the SCN at which the target standby became a primary database, and sets up the redo transport configuration from the new primary to the standby.

    If the bystander standbys are behind, then the broker simply sets up the redo transport configuration from the soon-to-be-new-primary to these standbys and completes the failover process.

  • NONE - During the failover process, the broker does NOT evaluate the status of the bystander standbys as part of the failover operation. They are marked as disabled with a status of ORA-16661 so that they can be evaluated later. The broker simply completes the failover to produce a new primary database as soon as possible.

    After the failover is completed, you can reinstate the bystander standbys. During reinstatement of a bystander, the broker determines whether the bystander is ahead of, or behind, the new primary. If the bystander is ahead of the new primary, then the broker automatically flashes back the standby to the SCN at which the target standby became a primary database and sets up redo transport from the new primary to the standby. (Even if a flashback is not required, the broker sets up the redo transport configuration from the new primary to these standbys.)

    The NONE option decreases the processing time for failover, but disables broker management of all bystander databases in the configuration. If fast-start failover is enabled, then the observer automatically reinstates the standby databases after failover has completed. Otherwise, you will have to manually reinstate the standby databases after failover has completed.