I have a situation whereby an Oracle Standby (managed by data guard) is really out of date. The redo logs are now being properly sync'd however there is a big gap.

I have the current SCNs for primary and standby but cannot find out how to get RMAN 9.2 to create an incremental backup from the specific SCN.

Does anyone have any experience with this?


3 Answers 3


I do not have access to a 9.2 database currently, but this is the method for doing this:

Using RMAN Incremental Backups to Roll Forward a Physical Standby Database

Basically, you can issue a BACKUP INCREMENTAL FROM SCN ... to create the desired backup. However I suspect this feature is not available in versions before, as there is no reference of it in the 9.2 Data Guard documentation nor the 9.2 BACKUP reference, and the related MOS notes also specify the lowest version as

Try the above syntax, but if it does not work, you will need to rebuild the standby.

  • thanks, no it doesn't. whilst you can find the scn, you can't backup from it Commented Oct 15, 2015 at 10:13

Try checking the alert log to figure out what sequence it needs next. If you have the archive logs online or in a backup online you can restore the missing archive log files and copy them over to the archive log directory to see if you can get the standby caught up. If you are running backups of your primary database, then just doing an incremental backup might not work because you still might miss changes from before the incremental backup. If you can't resolve the gap in archive logs you might need to restore the standby database again.

  • thanks, echoes @Balazs response. think this may be my only option Commented Oct 15, 2015 at 10:15

The incremental backup from SCN is a 10g feature.

From my point of view, you have 4 possible scenarios:

Scenario 1 - Regular Database with old archive logs

In this scenario you can restore the archives to the right location, then issue a recover operation using RMAN (connected to the standby database) and then start the standby recovery process

Scenario 2 - Regular Database without old archive logs

In this case, the faster bet is making a disaster recovery operation to "duplicate" the database.

Scenario 3 - Massive Database

In the case that you're working with a multi-terabyte database, you have to check what will take longer, getting the old logs (if any are available) to recover the standby or creating a "new" standby database.

  • I don't have all the redo logs, and cannot do incremental backup from scn. I think scenario 2 is my best option now Commented Oct 15, 2015 at 10:14
  • Good, then if you can, make a cold backup of the database. If not, do your regular RMAN backup and apply the archivelogs after the restore operation in the standby location. Configure the database as standby and you're done!
    – Silvarion
    Commented Oct 15, 2015 at 15:12
  • by a cold you mean down the primary and copy the data files? is it possible to do an RMAN backup and restore from a different physical location? Commented Oct 16, 2015 at 10:25
  • By cold backup I mean you put the database in the mount state and take and RMAN backup. Making those backups available to the standby database will allow you to restore it to the exact point when you took that backup.
    – Silvarion
    Commented Oct 16, 2015 at 12:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.