Re: Some Dataguard is good, lots more must be better?

  • From: Mladen Gogala <mgogala@xxxxxxxxxxx>
  • To: duncan.lawie@xxxxxxxxxxxxxxxxx
  • Date: Wed, 20 Sep 2006 04:51:18 -0400

On 09/20/2006 04:34:58 AM, Lawie, Duncan wrote:

> Say I want to push database 42 back by 2 hours 16 minutes and 42 seconds.  If 
> it is sharing a 
> filesystem with a bunch of other databases which are running just fine, how 
> do I load up the 
> old files for database 42, if the snapshot is at the filesystem level?  I'm 
> reading your mail 
> as saying the snapshot from time x would be a mountable filesystem?  If so, 
> do I then need to 
> copy the files back to my "real" filesystem, or do I change my database 
> control file to point 
> to the new mount?  The latter would appear to gradually fragment the few, 
> large filesystem 
> approach.  The former doesn't seem to save me time against some other 
> recovery mechanisms.

Duncan, Kevin is talking about NetApp which has old snapshots permanently 
mounted as 
read-only file system. Snapshots are not supported with RMAN (this may have 
changed since
the last time I worked with that) so that, in order for the database to be 
from snapshots, you would need to put your database in backup mode immediately 
before the 
snapshot is taken and take it out of the backup mode after the snapshot is 
taken. This works
pretty well. The mechanism to use for traveling back in time is, of course, 
database recovery.
Dr. Who is still a science fiction.

Mladen Gogala


Other related posts: