Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.server -> Re: ORA-00220: controlfile not mounted by first instance,
"Shravana Kumar" <kumar.shravana_at_blr.spcnl.co.in> wrote in message
news:3E13528F.8837ABFC_at_blr.spcnl.co.in...
> Hi All,
>
> We are using oracle 9.0.1 on Sun Solaris 8.
> I have the following problem while starting the database instance.
>
> SQL> startup
> ORACLE instance started.
>
> Total System Global Area 126804664 bytes
> Fixed Size 434872 bytes
> Variable Size 104857600 bytes
> Database Buffers 20971520 bytes
> Redo Buffers 540672 bytes
> ORA-00220: controlfile not mounted by first instance, check alert log
> for more
> info
> ORA-00202: controlfile: '/export/home/oracle/dev/SAMPLE/control02.ctl'
>
>
> The original control files ware deleted unfortunately. so i have to
> bring them back from backup. since then, i get this error.
>
> Any one knows, how to fix problem. thanks a lot,
>
Learn to visit http://tahiti.oracle.com : you'll find there a complete set of documentation for Oracle, including an error message search facility.
In your case, you'd see the following for error 220:
Cause: The specified control file has a different mount ID than the other control files that are being mounted. This means that the first instance to mount the database did not use this control file.
Action: Find and install the correct control file.
The action says it all, really. You've restored the wrong controlfiles, so find the correct ones and restore them instead.
And if you're going to then say you've definitely restored the right controlfiles, I'd say your database was stuffed (possibly you took a 'hot' copy of the controlfiles, rather than backing them up correctly with the appropriate 'alter database' command. Hot backups of anything in Oracle are likely to be internally inconsistent (ie, corrupt), so unless Oracle provided a mechanism for making them consistent again, they are a definite no-no. Redo makes hot datafile backup internally consistent, but redo isn't applied to controlfiles, and no other 'fix up' mechanism is provided either. So hot backups done simply by using O/S copy commands just don't work).
If your database is stuffed, blow it away and start again. From the number of posts you've made lately, I'm guessing you are fairly new to this Oracle malarkey, so hopefully blowing your database away and starting again won't be too much of a pain.
One way or another, you need to learn about controlfile multiplexing. The loss of all controlfiles is always a very expensive process in terms of downtime incurred, and multiplexing should mean you avoid the need for this.
You should also check out the 'alter database backup controlfile to trace' command, as that will enable you to re-create lost controlfiles without the need to restore them from a binary backup. This is the preferred mechanism for backing up controlfiles, incidentally, rather than taking binary backups.
Regards
HJR
> Best Regards,
> ShravanaKumar.
>
Received on Thu Jan 02 2003 - 03:34:33 CST
![]() |
![]() |