Many of my clients use storage backup (snapshots, etc.) to backup their databases. The right and supported way to do that is like any hot backup: place the database in backup mode, take the snapshot, take the database out of backup mode, archive the current logfile and backup the archives, right?
Some clients use the unsupported way of just taking the snapshot, as it is a consistent copy (exactly like a server crash). From my experience it works, but you can only restore the database to the snapshot time and you cannot apply archives.
Well, not any more! In Oracle 12c, as long as the snapshot mechanism meets several requirements (the database is crash consistent, write order is preserved and the snapshot stores the time of completion) you can take a snapshot without putting the database into backup mode. To perform the recovery, you will need to add the new “snapshot time” clause. Much easier than before!
Did You Know #13 – Storage Backup
Categories:
2 thoughts on “Did You Know #13 – Storage Backup”
Leave a Reply
Related Post
Did You Know #28 – Standby + RACDid You Know #28 – Standby + RAC
In an Oracle Data Guard configuration, the primary and standby databases can have different configurations. It’s very common to have a smaller server for the standby database (less CPUs, less
Did You Know #31 – OCR and Voting Disks in 18c and 19cDid You Know #31 – OCR and Voting Disks in 18c and 19c
In Oracle GI 18c, the OCR and Voting disks must be located on ASM. A certified NFS or 3rd party shared file system is not supported for these files (you
Installing PSU (GI+RAC+JVM)Installing PSU (GI+RAC+JVM)
I’ve installed quite a few PSUs in my professional life, but this time, it was something else. I’m talking about a RAC environment with 2 nodes and a database containing
Hi Liron,
Thanks for this post!
“From my experience it works, but you can only restore the database to the snapshot time and you cannot apply archives.
Well, not any more! In Oracle 12c, as long as the snapshot mechanism meets several requirements (the database is crash consistent, write order is preserved and the snapshot stores the time of completion)”
Actually, it was supported also before Oracle 12c. In my previous job (in the IDF) we used
Netapp crash consistent backups. See: http://www.netapp.com/us/media/tr-3858.pdf
“A crash-consistent image, just like a regular backup image, can be restored and rolled forward by
applying current online redo logs and archived logs.”
See also My Oracle Support (MOS) note 604683.1
Thanks for the comment Pini,
I haven’t tried that lately, and I don’t remember which version I had back then (10g? Maybe even 9i), but I definitely didn’t manage to roll archives forward…
Anyway, good to know.