From what I have read - the redo for a 12C database with PDB is at the root level. What that means to me is that there is one redo stream, and as the documentation put it, virtualized at the PDB level.
At a point in time T1 - lets assume I have 2 12C instances DB1 and DB2.
With a PDBA in DB1.
At Time T2 I take the PDBA from DB1 and plug it into DB2.
What happens to the recovery
A variety of possibilities
1. At Time T3 where T3 > T2 I need to restore the database DB1 prior to T2 and recover it upto T3
2. At Time T3 again T3 > T2 I need to restore DB2 from time prior to T2 and recover to point T3
Will we be able to take redo logs from DB1 and apply to DB2 for PBDA
Or - especially for point 2 will we need to restore DB1 and PDBA, recover till T2, then plug it back into DB2 and the apply recovery to DB2
You bring up good points. Currently the redo logs cannot be transported as part of the unplug/plug process.
1) You'll be able to restore DB1 prior to T2 and recover it upto T3. The dropped pdb's datafiles would simply not be part of this recovery session.
2) You'll not be able to restore DB2 prior to T2, since there will be no backups cataloged for the pdb plugged in at time T2. You could unplug the pdb again from DB2, perform the recovery and plug it back in after the recovery is done however.