A few months ago, I have come across of an interesting circumstances while doing an RMAN point-in-time-recovery (PITR) on one of our business critical RAC databases. I thought it would be worthy to blog about the interesting scenario that I come across. Ok, I need to take my memories back in time to explain what had happen.
At that time, I was advised to rollback the database using the last night ONLINE RMAN backup and recovery (forward) it to a particular point in time. I was able to restore the database successfully using the last night backup. However, while doing the recovery, RMAN was requesting archived logs that were generated almost a week back. I was surprised a little bit about this behavior as I was in a impression that RMAN looks for the archived logs that were generated after the online backup. Though I have restore and recover the database successfully to a PIT, I thought of conducting a mini quiz on this behavior to get cross your perspective on this, before I reveal the secrets and blog about it. So, here goes your quiz:
Under which circumstances does an RMAN recovery needs archive logs generated before the backup sets that are used in the restore operation?
Best of luck and awaits for your response....
Jaffar
12 comments:
If we have level 0 and level 1 backups, you need only archived redo logs that are generated after level 1 backup.
RMAN uses incremental backups during recovery instead of using archived redo log files, we need to backup all archived redo log files that are generated during backup procedure.
Old archivelogs could be needed if the backup pieces did not contain the full level 0 database but a subset of the datafiles.
Tycho
might be that, you took incremental backup by skipping 1 or more tablespaces
Hi Jaffar,
i think you have not restored controlfile which was generated at the time of backup which you are restoring..
The controlfile is may be older and datafiles are newer..Is it?
awaiting for your reply.. Thanks
Thanks everyone who participated in this mini quiz of RMAN. Very soon I would be going to reveal the secret.
Jaffar
some of the blocks in the restored backup is corrupted, hence it is asking for the old archived logs to perform block media recovery.
Syed Muzaffar Hussain.
seems you have set :-
"set until sequence" to very old which is not available in the curent RMAN archive log backup pieces..
thanks
anil
May be that particular tablespace was readonly and was skipped from getting backed up daily.
-Sai
Hi,
Maybe the issue was with the way that backup was taken the case could be like this:
The online backup was taken but without executing :
Alter system archive log current;
So there was active log files & database activity was low and/or size/number of redo log files was large.Thus it took time to the active redo log being archive .
KPRDBA
Jaffar,
can you share your thoughts on RMAN quiz? I'm curious to know.
Its quite possible that you may not have recoverd the database using the AUTOVACKUP of teh controlfile.
Post a Comment