one interesting case

dear folks:
following is one interesting case happened here last week. but I don't quite understand how the solution provided by Oracle Support works.
situation:
the user told me their Hard-disk crash and lost every thing!! therefore, we have to rebuild
the whole system from scratch. the only thing we have is hot-backup done last night,
which only include datafiles and control file. not thing else, no archived redo log, nothing!!

after re-install new Hard-disks, re-install LINUX, re-install Oracle DB Software, we done the
following follow Oracle Support's instructions:
1. restore the only thing we have: datafiles and control file.
2. startup mount;
3. recover database using backup controlfile until cancel;
4. alter database open resetlogs;
5. export full database;
6. import full database;

after those steps, our database seems back to normal and data is accessable to users!

my questions are:
1. what will Oracle do at step 3 ??
(i think Oracle will try to apply proper "archived redo log", but because archived was not backup,
Oracle will unable to recover the database. and at this time datafiles are not synchronized.
is this corrent??)
2. what will Oracle do at step 4 ??
(i think since Oracle requires synchronize to open the DB, Oracle will find out the datafile
will the smallest SCN and reset all database files to this SCN. therefore, after step 4
the database is available to users and the data will back to the beginning of our last night's
hot-back. is this correct ??)
3. is step 5 and 6 mandatory in this recovery action ??
(i think this recovery action finish at step 4. export/import will not effect the synchronize of
the DB. we just use it as a post-recovery full backup)


thank you very very much for your kindly answers...

wayne

what you need is read backup & recovery concepts, i.e RTFM

obviously you did something after step 3, what was it?

Dear pando,stmontgo and all other folks:

following are all the steps I did before and after the step 3 "Recover" command,but i'm very confused about how this works. So, any answer to my previous Questions are very appreciated:

startup mount;
recover database using backup controlfile until cancel;
alter database open resetlogs;

(I think we got the following error and then add 4 lines in the init.ora AFTER the alter database command not BEFORE it,but i'm not every sure about this)
got ora-3113 => check alert file:
ora-600 internal error [4194] [29] [29] ,

add the following 4 lines in the init.ora:
_allow_resetlogs_corruption = true
_corrupted_rollback_segments = (rbs1, rbs2, rbs3)
event="10269 trace name context forever, level 10"
event="10061 trace name context forever, level 10"

retry:alter database open resetlogs;
DB opened.

1 corrention about my previous reply:
AFTER adding that 4 lines in the init.ora, we did:
shutdown immediate;
startup mount;
recover database using backup controlfile until cancel;
alter database open resetlogs;
DB opened.

Thanks for your time~