Re: Control file recovery

On 5 Sep 2005 03:49:19 -0700, "Pradeep" <agarwalp_at_eeism.com> wrote:

>Status of Status of Datafiles Response
>Online Logs
>
>1. Available Current If the online logs contain
> redo necessary for recovery, then
> restore a backup control file
> apply the logs during recovery.
> Hence, you must specify the
> filename of the online logs
> containing the changes in order to
> open the database. After recovery,
> open RESETLOGS.
>
>2. Unavailable Current If the online logs contain redo
> necessary for recovery, then
> you must re-create the control
> file. Because the logs are
> inaccessible, open RESETLOGS.
>
>3. Available Backup Restore a backup control file,
> perform complete recovery,
> and then open RESETLOGS.
>
>4. Unavailable Backup Restore a backup
> control file, perform incomplete
> recovery, and then open RESETLOGS.
>
>Above is the extract from Oracle User Managed recovery Guide, but not
>quite understood the 1st and 2nd points. May be i am missing a samll
>thing
>
>Q1. In point 1 and also in point 3 , why do we have to do use Reset
>logs, even though we have the datafile and Online redo logs ?
>Q2. In point 2 it says "since the logs are inaccessible re-create the
>control file". Why do we have to create a new file? Why can't we use a
>backup control file and do incomplete recovery as in point 4?

Boils down to the same issue.
Every datafile has the current SCN and the sequence number of the logfile.
That information is also in the controlfile and in the online redolog. If you don't have the online redolog how should Oracle synchronize those bit of information during database open? Why should Oracle trust the information in the datafiles is correct? It can't trust the information, because you don't have the affected redolog files anymore. So your chain of recovery is broken anyway. That is why it needs to perform resetlogs, ie reset the log sequence number to 1 (the SCN is unaffected).  

--
Sybrand Bakker, Senior Oracle DBA