I stumbled upon a very strange concept which comes with autobackup of the controlfile. The point I feel very bad about is that the controlfile backup happens AFTER the backup activity which triggered it. So it happens that the autobackup of the controlfile has a higher SCN than all the stuff which has been backuped right before it.
First a little inside into my setup:
- rman full/incr. backups are done each night and so creating a controlfile autobackup
- rman archivelog backups are done each hour and so creating a controlfile autobackup
- backups are stored on completly separate storage infrastructure then datafiles and FRA
So if I want to clone my database using rman-backups I cannot use the most recent autobackup controlfile with the associated archivelog backup as it will end in an inconsistent recovery due to missing archivelogs with required higher SCN. That is kind of OK for the purpose of cloning as I could use some older controlfile. But in case I loose the whole database and I am only left with my backup area I cannot return to that very most current controlfile (and so the very last state of the database).
Is there anything I can do to make each (so including the very most current) set of controlfile autobackup and archivelog backup to be fully recoverable?
Or is there something wrong in the way I see those things?
Any other ideas to work around it?
Regards,
Robert
-------------
Database 12.1.0.2