Scenario 1: Prepare the Databases

Most of the usual Data Guard preparation steps apply when setting up a DG PDB environment. Notable differences are that standby redo logfiles (SRLs) are added later at the PDB level after the Data Guard DG PDB configuration has been created (instead of beforehand at the CDB level) and that CDB-level file name conversion initialization parameters are not used for DG PDB configurations.

For the source and target databases, use SQL*Plus to connect AS SYSDBA and perform the following actions (if they have not been already) to support future PDB-level role transitions:

Enable force logging.
SQL> ALTER DATABASE FORCE LOGGING; 
Database altered.
Enable flashback database.
SQL> ALTER DATABASE FLASHBACK ON; 
Database altered.
If necessary, update the initialization parameters in the server parameter files for each database.

The CDB-level file name conversion initialization parameters db_file_name_convert and log_file_name_convert are not used for DG PDB configurations, but the dg_broker_start, log_archive_dest_1, and standby_file_management parameters should be set to the values as show in the table below. In addition, if multi-instance databases are used, the dg_broker_config_file1 and dg_broker_config_file2 initialization parameters must be updated to specify locations accessible by all instances. Because the source and target container databases do not share the same db_name value, including the db_unique_name parameter in the spfile is optional for DG PDB configurations.

Configure the following initialization parameters on the source database (using scope=both):

Configure the following initialization parameters on the target database:

DGMGRL must connect to the source and target container databases using an account with SYSDBA privileges to create and manage a DG PDB configuration. For the scenarios, the existing SYS user for each database will be used for this purpose. Optionally, a separate user can be created on each database and granted SYSDBA privileges. Note that because the source and target databases are not copies of each other, the passwords for the management accounts do not need to match on the source and target databases and the password file from the source database does not need to be copied to the target database as is required for CDB-level Data Guard protection.