Re: help with recovery

"Howard J. Rogers" <hjr_at_dizwell.com> wrote in message news:<4060d1f0$0$3956$afc38c87_at_news.optusnet.com.au>...
> "daveb" <davebest_at_SuPsAaM.net> wrote in message
> news:EtOdnRJrCONcV_3dRVn-tw_at_speakeasy.net...
> > We have a small test database (9.2.0.1 on Windows 2000) and the hard-drive
> > just fried. Should be no biggie since we have create scripts for the
> > schema. But now we find we need some of the test data...no db backup on
> > that, d'oh. But it turns out that we have a copy of the oracle directory.
>
> Strangely enough, some degree of precision is required when describing a
> problem before a solution can usefully be offered. What oracle directory? Do
> you mean the "oradata" directory? If so, what -exactly- is in that
> directory? Are there control files? Are there things which look like redo
> logs? Are there all the data files of the entire database before the hard
> disk failure?
>
> You know, just a few minor details would help!
>
> > The question is how to access it.
>
> The question is, what is there to access?? See above.
>
> >I did a fresh Oracle install on another
> > machine
>
> Good.
>
> >and created a database of the same name,
>
> Oh dear. Why do people insist on creating new databases when their actual
> objective is to recover the old one?

Because until they are taught otherwise, they think that a database consists of data and executables, so you should be able to plug some existing data files into a new database... right? Recovery as a specific set of procedures with specific requirements resolving a constantly fuzzy stream of data is not obvious in the larger scheme of things.

So the first issue is one of semantics - defining recovery, defining a consistent database. Oughta be a FAQ.

jg

--
@home.com is bogus.
Happy 93rd Birthday Joseph Barbera. 
http://www.clampettstudio.com/bios/joseph_barbera.htm
We'll ignore the smurfs...