Which three statements are true about using flashback database in a multitenant container database (CDB)?

Flashing back a CDB affects all containers - the root and all associated PDBs are returned to the target point together, because the CDB shares a single set of redo and flashback logs across all containers.

The ALTER DATABASE FLASHBACK ON command used to enable flashback on a CDB must be issued while the database is in MOUNT state, because the operation modifies control file settings and requires exclusive access before the database is opened.

Flashing back an individual PDB independently requires local undo mode to be enabled on the CDB and the use of a PDB-level restore point - it is not a universally available capability in all CDB configurations.

The DB_FLASHBACK_RETENTION_TARGET parameter must be set to a non-zero value to define how many minutes of flashback logs the system should retain, and this is a required prerequisite for enabling CDB-level flashback database.

In CDB flashback operations, the supported target specifications are a timestamp or an SCN - restore point-based flashback in a CDB environment has behavioral restrictions that distinguish it from non-CDB flashback usage.