Recovery in sqlplus

Question

Hi;
When I shutdown abort, instance recovery is done automatically once I open the database.
What I want to ask is, when I issue
'recover database using backup controlfile' statement even only once,
I can not recover the database with normal 'recover database' clause.
Does this statement modify controlfile or datafiles? Whats the reason for this?
Here is the demo:
SQL>shutdown abort;
ORACLE instance shut down.
startup mount;
ORACLE instance started.
Total System Global Area 285212672 bytes
Fixed Size 1287016 bytes
Variable Size 92277912 bytes
Database Buffers 188743680 bytes
Redo Buffers 2904064 bytes
Database mounted.
SQL>recover database using backup controlfile;
ORA-00279: change 3574972 generated at 12/06/2009 18:34:54 needed for thread 1
ORA-00289: suggestion :
C:\ORACLEXE\APP\ORACLE\FLASH_RECOVERY_AREA\XE\ARCHIVELOG\2009_12_06\O1_MF_1_4_%U_.
ORA-00280: change 3574972 for thread 1 is in sequence #4
Specify log: {suggested | filename | AUTO | CANCEL}
cancel
Media recovery cancelled.
SQL>recover database;
ORA-00283: recovery session canceled due to errors
ORA-01610: recovery using the BACKUP CONTROLFILE option must be doneIf I issue recover database at first rather than 'recover database using backup controlfile', I can successfully do the recovery.

Answer 1

Hi john,
Let me try to explain, See, basic thing, usually we go for "using backup control file" on conditions
- Controlfile is older than any of your datafileson current DB at any moment, since all the c/r/d file mantain consistency.
- worst case when valid current control file is corrupted or deleted.
Coming, to your Issue, from your demo, if you see you have mounted the DB - at this moment it's Check the controlfile (reads) and checks the rest of the data and redo logs for validity - consistency Checks will be carried and logs will be ready for new transactions. At this moment if you are carry out the recovery where you are using the old control file, which does not know the new changes carried out and you tried to carryout the recovery - as a part of it - it inturn generated Redo Vectors - @ that point of moment if you check the redo vectors all are in process of current DB and did not expected for recovery. since due to recovery they got mixed up( or got updated with in sync old control file). Now if recovery continues..but it should stop until the point of all archives are applied and .. it must the control file and data file for consistent state to stop the recovery.
Since, we need to for provided options, either cancel or provided available logs till the scn's are in sync. Due to that you need to startup the DB with New Reset Logs.
HTH
- Pavan Kumar N
Oracle 9i/10g - OCP
http://oracleinternals.blogspot.com/