导言：如果你的数据库昨天下午5点做全了备份，昨天晚上7点为表空间增加了文件。今天早上数据库磁盘坏了。 请问：

1，数据库能够完全恢复吗？

2，如何恢复？

我回答：应该可以，但是具体步骤吗。。。。这个当时真的头晕了。。没有回答出来。 惭愧呀。。。。

实验环境：

WINDOWS+ ORACLE9.2.0.1,

操作步骤：

1, rman backup full database ;

2, rman backup current control file ;

3, sql> create tablespace ts_lost datafile size 10m;

4, sql> create table t1 tablespace ts_lost as ..... ;

 sql> create table t2 tablespace ts_lost as ..... ;

alter system switch logfile ;

5, shutdown immediate ;

6, delete all database files (control files, dtafiles) from disk, exclude redolog file .

-----------------OK, now let's see how to completely recover the database .

step :

1, sql> startup nomount

2,rman> restore controlfile from "ctl.bak" ;

3,

# rman> restore database ;  --- get error. because file "ts_lost " not backuped.

RMAN-03002: failure of restore command at 05/11/2010 18:48:27
RMAN-06026: some targets not found - aborting restore
RMAN-06023: no backup or copy of datafile 12 found to restore

But I have a question. How  oracle know the newest datafile with the old backup controlfile. Maybe It look at file from system tablespace

sql> select file# from v$recover_file ; -- assume 1,2,3,4,5,6,7

rman> restore datafile 1 ;

rman> restore datafile 2 ;

rman> restore datafile 3 ;

...........

4,

alter database create datafile 12 as 'C:\oracle\oradata\mydb\ts_test.dbf' ;

5, recover database using backup controlfile;

19:12:05 SQL> recover database using backup controlfile ;
ORA-00279: change 2098519 generated at 05/11/2010 17:22:11 needed for thread 1
ORA-00289: suggestion : C:\ORACLE\ORADATA\ARC00034.001
ORA-00280: change 2098519 for thread 1 is in sequence #34

19:12:13 Specify log: {=suggested | filename | AUTO | CANCEL}

ORA-00279: change 2098606 generated at 05/11/2010 17:24:46 needed for thread 1
ORA-00289: suggestion : C:\ORACLE\ORADATA\ARC00035.001
ORA-00280: change 2098606 for thread 1 is in sequence #35
ORA-00278: log file 'C:\ORACLE\ORADATA\ARC00034.001' no longer needed for this recovery

19:12:20 Specify log: {=suggested | filename | AUTO | CANCEL}
C:\oracle\oradata\mydb\redo01.log
Log applied.
Media recovery complete.

--note, you must sure the archive log and the latest online redo log pathname.

If oracle prompt a not exist archived log to apply, then it means to apply an online redo logfile . Then should manually input the redolog pathname .

6, alter database open resetlogs;  -- because we use backup controlfile to restore and recover database, so must use resetlogs to open db.

memo:

1, only recover database can use "using backup controlfile" caluse

19:11:39 SQL> recover datafile 12 using backup controlfile ;
ORA-00274: illegal recovery option USING