Propagating Unrecoverable Operations to Standby Databases Oracle Database Tips by Donald BurlesonDecember 9, 2015

 Propagating Unrecoverable Operations to Standby Databases

The standby_file_management initialization parameter plays an important role in the recovery process when the attributes of the data files are modified on the primary site. For the ease of administration, this parameter should be set to AUTO in the init.ora file on the standby site.

            TIP - If the Oracle instance is created on raw device, STANDBY_FILE_MANAGEMENT should be set to MANUAL.

When this is set to AUTO, adding a tablespace or datafile on the primary database will automatically propagate to the standby database. As a part of the recovery, Oracle will create the required operating system files.

The following section presents details of events that require intervention of the DBA and the possible solution to reflect these changes to the standby site.

Changes in the Tablespaces or Datafiles

In this section, information on the steps required to update the Oracle instance when the tablespace or datafiles on the primary database are changed and the automatic file management is not enabled on the standby database. It is assumed that the Oracle instance is running in managed recovery mode.

Creating Tablespace and Datafiles in the Primary Database

If a tablespace or a datafile has been added to an existing tablespace in the primary database, the following steps must be performed in order to synchronize the standby site. The following steps should be performed on each physical standby site:

• After altering the tablespace on the primary site, take it offline:

ALTER TABLESPACE <tablespace name> OFFLINE;

• If it is a new tablespace, copy all the datafiles associated with this tablespace to the standby site. If datafiles have been added to an existing tablespace, only copy the new datafiles.
   

• Bring the tablespace online on the primary database:

ALTER TABLESPACE <tablespace name> ONLINE;

• Switch a logfile on the primary database so the data dictionary on the standby site will be updated by the managed recovery process:

ALTER SYSTEM SWITCH LOGFILE;

The new datafiles can be verified from the v$datafile view on the standby database.