Applying redo - recovery fundamentals 2005-11-27 - By Mark Bole
orcl wrote:
> All, Im having big trouble in a recovery scenario, here is what I have
>
> 1, A small shell of a database with all datafiles copied to a backup
> location (cold backup)
> 2. Database is in archivelog mode
> 3. startup up the database and create a table with 6K records. I can
> see all the redo getting generated wait a few min and drop the table .
> redo logs are 5mb
> 4. Shutdown the database, delete all datafiles and copy over all
> database files (from step 1) including redo logs
> 5. Startup mount and issue recover database until cancel
> 6. I *dont * get prompted for cancel and in 2 seconds i get "media
> recovery complete" and my table *is not* restored
>
> Basically the same problem if I use recover database until time
> '2005-11-27:09:25:00';
>
> What the heck am I doing wrong?
>
[...]
You need to either restore your current controlfile (not the one you
previously backed up), or create a new one based on the contents of
"backup controfile to trace". If you simply restore a closed consistent
whole database backup, the instance will correctly determine that no
recovery is needed.
Also, it's not clear in step 4 whether you are referring to archived
redo logs or online redo logs. It is strongly recommended (by Oracle)
to never backup (or restore) online redo logs.
I personally find that practice with manual recovery is a good thing,
but you can also use RMAN to handle many of the details for you.
--
Mark Bole
http://www.bincomputing.com
--
http://www.freelists.org/webpage/oracle-l
|
|
