Re: howto drop a corrupt database ?

On Mon, 10 Nov 2003 20:49:59 GMT, Brian Peasland <dba_at_remove_spam.peasland.com> wrote:

>> >It is said that 10g will provide a drop database command. Though I'd believe that
>> >the datafiles need to be OMF for oracle to clean up the datafiles also.
>> >
>> >Anurag
>> >
>> Why would they have to be OMF? Oracle has always been able to create
>> OS files (CREATE TABLESPACE . . .), so why couldn't it delete them?
>
>Probably the same reason a simple DROP TABLESPACE command can't delete
>the datafiles. For some reason (?) a simple DROP TABLESPACE won't delete
>the datafiles unless the files are OMF. There is a clause (INCLUDING
>CONTENTS AND DATAFILES) that will do it as well. Maybe DROP DATABASE in
>10G has a similar clause.
>
>Cheers,
>Brian

Then back to the statement that started this branch of the thread (that a DROP DATABASE command would require OMF to be able to clean up the data files) . . . I guess it depends on what the meaning of "is" is .. ;-)

I took that statement to mean that it was thought that Oracle is inherently unable to drop OS files, thus the developers would not be *able* to implement dropping of OS files. I guess others read it to mean that the developers would *choose* not to implement dropping of OS files, thus making OMF a prerequisite to using a proposed DROP DATABASE command.

It should be obvious that Oracle has the ability to both create and drop OS files. Now, how the developers choose to make use of that with any particular command (CREATE | DROP TABLESPACE, CREATE | DROP DATABASE) is another matter. I take it as a matter of "can't" vs. "won't", but I guess if "won't" is implemented in the commands, the difference becomes a moot point. Received on Wed Nov 12 2003 - 09:29:32 CST