使用备份控制文件恢复数据库

    有备份的control file，但之后我做了ddl操作，导致当前的control file比备份的新，然后control file全部丢失，怎么用备份的control file来恢复数据库？

    步骤1：检查当前redo和archive文件

        sys@PAN>select group#,sequence#,archived,status,first_change# from v$log;

 

        GROUP#  SEQUENCE# ARC STATUS           FIRST_CHANGE#

    ---------- ---------- --- ---------------- -------------

             1          4 NO  CURRENT                 443342

             2          2 YES INACTIVE                443335

             3          3 YES INACTIVE                443338

            

        sys@PAN>select name from v$archived_log where name is not null;

 

        NAME

        --------------------------------------------------------------------------------

        /u01/flash_recovery_area/PAN/archivelog/2013_06_23/o1_mf_1_1_8wg0nhky_.arc

        /u01/flash_recovery_area/PAN/archivelog/2013_06_23/o1_mf_1_2_8wg0njog_.arc

        /u01/flash_recovery_area/PAN/archivelog/2013_06_23/o1_mf_1_3_8wg0nlhz_.arc

    步骤2：冷备

    步骤3：

        alter database backup controlfile to trace;

        生成一个trace文件，可以用于恢复控制文件，它是一个文本文件。

    步骤4：

        alter database backup controlfile to '/backup/pancontrol.bak';

        备份当前的控制文件。

    步骤5：

        创建一个表空间，然后在此表空间里建表，加数据，目的是让controlfile改变成新的。

        create tablespace ts_test02 datafile '/u01/oradata/pan/ts_test02_01.dbf' size 10m;

        sqlplus scott/tiger

        create table tb_test02 (id int) tablespace ts_test02;

        insert into tb_test02 values(1);

        insert into tb_test02 values(2);

        insert into tb_test02 values(3);

        commit;

        现在这些改变在哪个日志文件里呢？

        sys@PAN>select group#,sequence#,archived,status from v$log;

 

            GROUP#  SEQUENCE# ARC STATUS

        ---------- ---------- --- ----------------

                 1          4 NO  CURRENT

                 2          2 YES INACTIVE

                 3          3 YES INACTIVE

 

        Elapsed: 00:00:00.01

        理论上在1号日志中，一会要用这个做恢复。

    步骤6

        删除当前controlfile

    步骤7

        shutdown abort

    步骤8

        用备份controlfile转储。

        [oracle@oracle11g pan]$ cp /backup/pancontrol.bak control01.ctl

        [oracle@oracle11g pan]$ cp /backup/pancontrol.bak control02.ctl

    步骤9

        起库到mount，若到open会出错：

        sys@PAN>alter database open;

        alter database open

        *

        ERRORat line 1:

        ORA-01589: must use RESETLOGS or NORESETLOGS option for database open

    步骤10

        尝试使用备份的controlfile来恢复数据库。

        recover database using backup controlfile;

        结果：

        sys@PAN>recover database using backup controlfile;

        ORA-00279: change 443470 generated at 06/23/2013 10:23:49 needed for thread 1

        ORA-00289: suggestion : /u01/flash_recovery_area/PAN/archivelog/2013_06_23/o1_mf_1_4_%u_.arc

        ORA-00280: change 443470 for thread 1 is in sequence #4

        Specify log: {<RET>=suggested | filename | AUTO | CANCEL}

       

        看下这个目录下有没有这个归档日志，发现没有。也就是说数据库会先找你的归档日志来恢复，但我们知道以前的那些DDL操作被记录到1号联机日志里了，所以我要用1号redo日志来做恢复。

       

        继续：

        recover database using backup controlfile;

        ORA-00279: change 443470 generated at 06/23/2013 10:23:49 needed for thread 1

        ORA-00289: suggestion : /u01/flash_recovery_area/PAN/archivelog/2013_06_23/o1_mf_1_4_%u_.arc

        ORA-00280: change 443470 for thread 1 is in sequence #4

        Specifylog: {<RET>=suggested | filename | AUTO | CANCEL}

        /u01/oradata/pan/redo01.log

        ORA-00283: recovery session canceled due to errors

        ORA-01244: unnamed datafile(s) added to control file by media recovery

        ORA-01110: data file 9: '/u01/oradata/pan/ts_test02_01.dbf'

        ORA-01112: media recovery not started

        又出错了！它说没有这个文件，因为日志里记录着ts_test02_01.dbf，但控制文件里却没有，好的，现在我查看下控制文件里那个文件叫什么。

        sys@PAN>select file#,status,name from v$datafile;

 

             FILE# STATUS NAME

        ---------- ---------------------------------------------------------------------------------------

                 1 SYSTEM  /u01/oradata/pan/system01.dbf

                 2 ONLINE  /u01/oradata/pan/sysaux01.dbf

                 3 ONLINE  /u01/oradata/pan/lxtbs01.dbf

                 4 ONLINE  /u01/oradata/pan/users01.dbf

                 5 ONLINE  /u01/oradata/pan/undotbs02.dbf

                 6 ONLINE  /u01/oradata/pan/ts_test01.dbf

                 7 ONLINE  /u01/oradata/pan/ts_readonly01.dbf

                 8 ONLINE  /u01/oradata/pan/app2_01.dbf

                 9 RECOVER /u01/oracle/dbs/UNNAMED00009

 

        9 rows selected.

        现在明白了，控制文件里的这个文件名为“/u01/oracle/dbs/UNNAMED00009”，我把它先重命名成正确的。

        sys@PAN>alter database rename file '/u01/oracle/dbs/UNNAMED00009'

        2  to '/u01/oradata/pan/ts_test02_01.dbf';

        现在再一次恢复，使用1号redo日志。

        recover database using backup controlfile;

        ORA-00279: change 443687 generated at 06/23/2013 10:27:38 needed for thread 1

        ORA-00289: suggestion : /u01/flash_recovery_area/PAN/archivelog/2013_06_23/o1_mf_1_4_%u_.arc

        ORA-00280: change 443687 for thread 1 is in sequence #4

        Specify log: {<RET>=suggested | filename | AUTO | CANCEL}

        /u01/oradata/pan/redo01.log

        Log applied.

        Media recovery complete.

        果然，恢复成功！

        步骤11

        alter database open resetlogs;

        步骤12 冷备