Re: RMAN backup piece validation

From: Dragutin Jastrebic <>
Date: Wed, 25 Sep 2013 22:59:53 +0200
Message-ID: <>

Well, one thing you can be sure of in IT , (as in life...) is that you can't be sure of anything in IT.
So, can the validity of your backup pieces be checked up to 100% , just with validate commands ?

My answer is no. It can't.

All the validate commands, (restore validate, backup validate, 11g has introduced the recover trial and so on)
are great tools, and you should use them to validate your backupsets.

Ideally you should have a job that is performing this validation on regular basis.

But this does not mean that you never need to test the real restore & recover , for exemple
by copying the production database in the test environment and then proceed.

You can see the "validate" operations as additional tools that make your life easier, but they are not a substitute for real tests.


Because there are always bugs in the code ,so perhaps that restore validate contains bugs as well.

There is no guarantee that the write phase will perform well, even when the read phase of restore says that there is no problem with the backupset.

I've already faced the situations when restore validate commands did not report any errors but
the real restore command did.

All projects work under budget constraints and sometimes it's not easy to convince your management to
give you a free server just to test your backup and recovery procedures but it is crucial to check
the validity of your backups and your restore strategies as well.

The good strategy that I've seen on several projects is to do daily (or weekly) cloning of the production database to some test platform. The database is then given to end users that don't need real time access of production data.

This is a very nice thing because

-The users can access the production data without disturbing the production database
-The production backup is validated

HTH Dragutin


Received on Wed Sep 25 2013 - 22:59:53 CEST

Original text of this message