WEB开发网
开发学院数据库MSSQL Server Sql数据库MDF数据文件数据库恢复 阅读

Sql数据库MDF数据文件数据库恢复

 2007-11-11 05:54:04 来源:WEB开发网   
核心提示: 3. 所有其他文件备份后的归档都在存在 恢复步骤: 1,拷贝任一未丢失文件的备份回来 2.创建新控制文件,但控制文件不包括丢失的数据文件 3.进行恢复 4.碰到 ORA-00283: recovery session canceled due to errors ORA-01244: unname

   3.     所有其他文件备份后的归档都在存在

   恢复步骤:

   1,拷贝任一未丢失文件的备份回来

   2.创建新控制文件,但控制文件不包括丢失的数据文件

   3.进行恢复

   4.碰到

   ORA-00283: recovery session canceled due to errors
   ORA-01244: unnamed datafile(s) added to controlfile by media recovery
   ORA-01110: data file 3: 'D:\Oracle(大型网站数据库平台)\ORADATA\xxxx\xxxx.DBF'

   5.select name from v$datafile找出uname file

   6. alter database create datafile 'D:\Oracle(大型网站数据库平台)\ORA92\DATABASE\UNNAMEDxxxxx' as 'D:\Oracle(大型网站数据库平台)\ORADATA\ xxxx\xxxx.DBF ';

   7.继续恢复

   8.恢复完成,打开.

   可能会遇到的问题:

   1.如果没有拷贝备份回来直接重建不包含丢失文件的控制文件时,可以打开数据库,但丢失的文件会显示成missingxxxx的file name,这时候就会不能恢复.

   试验过程:

   Microsoft Windows XP [版本 5.1.2600]

   (C) 版权所有 1985-2001 Microsoft Corp.

   C:\Documents and Settings\qigong>oradim -startup -sid test

   C:\Documents and Settings\qigong>sqlplus / as sysdba

   SQL*Plus: Release 10.1.0.2.0 - Production on 星期三 8月 25 15:22:39 2004

   Copyright (c) 1982, 2004, Oracle(大型网站数据库平台).  All rights reserved.

   Connected to:
   Oracle(大型网站数据库平台) Database 10g Enterprise Edition Release 10.1.0.2.0 - Production
   With the Partitioning, OLAP and Data Mining options

   SQL> shutdown immediate;
   Database closed.
   Database dismounted.
   Oracle(大型网站数据库平台) instance shut down.
   SQL> exit
   Disconnected from Oracle(大型网站数据库平台) Database 10g Enterprise Edition Release 10.1.0.2.0 - Producti
   With the Partitioning, OLAP and Data Mining options

   C:\Documents and Settings\qigong>sqlplus / as sysdba

   SQL*Plus: Release 10.1.0.2.0 - Production on 星期三 8月 25 15:25:37 2004

   Copyright (c) 1982, 2004, Oracle(大型网站数据库平台).  All rights reserved.

   Connected to an idle instance.

   SQL> startup nomount;
   Oracle(大型网站数据库平台) instance started.

   Total System Global Area  163577856 bytes
   Fixed Size          787948 bytes
   Variable Size        74447380 bytes
   Database Buffers      88080384 bytes
   Redo Buffers         262144 bytes
   SQL> alter database mount;
   alter database mount
   *
   ERROR at line 1:
   ORA-00205: error in identifying controlfile, check alert log for more info

   SQL> shutdown
   ORA-01507: database not mounted

   Oracle(大型网站数据库平台) instance shut down.
   SQL> startup
   Oracle(大型网站数据库平台) instance started.

   Total System Global Area  163577856 bytes
   Fixed Size          787948 bytes
   Variable Size        74447380 bytes
   Database Buffers      88080384 bytes
   Redo Buffers         262144 bytes
   Database mounted.
   Database opened.
   SQL> alter system switch logfile;

   System altered.

   SQL> /

   System altered.

   SQL> /

   System altered.

   SQL> create tablespace test2
   2  datafile  'c:\test2.dbf' size 10m
   3      extent management local
   4      segment space management auto
   5      uniform size 100k;

   Tablespace created.

   SQL> insert into testlost values(2);

   1 row created.

   SQL> commit;

   Commit complete.

   SQL> create table testlost2(a number) tablespace test2;

   Table created.

   SQL> shutdown immediate;
   Database closed.
   Database dismounted.
   Oracle(大型网站数据库平台) instance shut down.
   SQL> startup nomount
   Oracle(大型网站数据库平台) instance started.

   Total System Global Area  163577856 bytes
   Fixed Size          787948 bytes
   Variable Size        74447380 bytes
   Database Buffers      88080384 bytes
   Redo Buffers         262144 bytes

   C:\Documents and Settings\qigong>rman

   Recovery Manager: Release 10.1.0.2.0 - Production

   Copyright (c) 1995, 2004, Oracle(大型网站数据库平台).  All rights reserved.

   RMAN> connect target

   connected to target database: test (not mounted)

   RMAN> restore controlfile from autobackup;

   Starting restore at 25-8月 -04
   using target database controlfile instead of recovery catalog
   allocated channel: ORA_DISK_1
   channel ORA_DISK_1: sid=159 devtype=DISK

   recovery area destination: C:\Oracle(大型网站数据库平台)\product\10.1.0\flash_recovery_area
   database name (or lock name space) used for search: TEST
   channel ORA_DISK_1: autobackup found in the recovery area
   channel ORA_DISK_1: autobackup found: C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0   \FLASH_RECOVERY_AREA\TEST\AUTOBACKUP\2004_08_24\O1_MF_S_535041963_0LOS1FCX_.BKP
   channel ORA_DISK_1: controlfile restore from autobackup complete
   output filename=C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\CONTROL01.CTL
   output filename=C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\CONTROL02.CTL
   output filename=C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\CONTROL03.CTL
   Finished restore at 25-8月 -04

   RMAN> exit

   RMAN-06900: WARNING: unable to generate V$RMAN_STATUS or V$RMAN_OUTPUT row
   RMAN-06901: WARNING: disabling update of the V$RMAN_STATUS and V$RMAN_OUTPUT rows

   SQL> alter database mount;
   alter database mount
   *
   ERROR at line 1:
   ORA-01991: invalid password file
   'C:\Oracle(大型网站数据库平台)\product\10.1.0\Db_4\DATABASE\PWDtest.ORA'

   SQL> DECLARE
   2  devtype varchar2(256);
   3  done boolean;
   4  BEGIN
   5  devtype:=sys.dbms_backup_restore.deviceAllocate (type=>'',ident=>'');
   6  sys.dbms_backup_restore.restoreSetDatafile;
   7  sys.dbms_backup_restore.restoreDatafileTo(dfnumber=>14,toname=>'c:\test1.DBF');
   8  sys.dbms_backup_restore.restoreBackupPiece(done=>done,handle=>'C:\02FUAT0K_1_1_2.
   9  sys.dbms_backup_restore.deviceDeallocate;
   10  END;
   11  /

   PL/SQL procedure successfully completed.

   SQL> shutdown immediate;
   ORA-01109: database not open

   Database dismounted.
   Oracle(大型网站数据库平台) instance shut down.
   SQL> startup nomount
   Oracle(大型网站数据库平台) instance started.

   Total System Global Area  163577856 bytes
   Fixed Size          787948 bytes
   Variable Size        74447380 bytes
   Database Buffers      88080384 bytes
   Redo Buffers         262144 bytes
   SQL> CREATE CONTROLFILE REUSE DATABASE "TEST" NORESETLOGS  ARCHIVELOG
   2    MAXLOGFILES 16
   3    MAXLOGMEMBERS 3
   4    MAXDATAFILES 100
   5    MAXINSTANCES 8
   6    MAXLOGHISTORY 454
   7  LOGFILE
   8   GROUP 1 'C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\REDO01.LOG'  SIZE 10M,
   9   GROUP 2 'C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\REDO02.LOG'  SIZE 10M,
   10   GROUP 3 'C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\REDO03.LOG'  SIZE 10M
   11  -- STANDBY LOGFILE
   12  DATAFILE
   13   'C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\SYSTEM01.DBF',
   14   'C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\UNDOTBS01.DBF',
   15   'C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\SYSAUX01.DBF',
   16   'C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\USERS01.DBF',
   17   'C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\EXAMPLE01.DBF',
   18   'C:\TBS1.DBF',
   19   'C:\TBS2.DBF',
   20   'C:\TBS3.DBF',
   21   'C:\TBS4.DBF',
   22   'C:\TBS5.DBF',
   23   'C:\TBS6.DBF',
   24   'C:\TBS7.DBF',
   25   'C:\TBS8.DBF',
   26   'C:\TEST1.DBF'
   27  CHARACTER SET ZHS16GBK
   28  ;

   Control file created.

   SQL> recover database;
   ORA-00279: change 1196994 generated at 08/25/2004 15:32:36 needed for thread 1
   ORA-00289: suggestion : C:\Oracle(大型网站数据库平台)\ARCHIVE\ARC00010_0535038675.001
   ORA-00280: change 1196994 for thread 1 is in sequence #10

   Specify log: {<RET>=suggested | filename | AUTO | CANCEL}
   auto
   ORA-00283: recovery session canceled due to errors
   ORA-01244: unnamed datafile(s) added to controlfile by media recovery
   ORA-01110: data file 15: 'C:\TEST2.DBF'

   ORA-01112: media recovery not started

   SQL> select name from v$datafile;

   NAME
   --------------------------------------------------------------------------------
   C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\SYSTEM01.DBF
   C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\UNDOTBS01.DBF
   C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\SYSAUX01.DBF
   C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\USERS01.DBF
   C:\Oracle(大型网站数据库平台)\PRODUCT\10.1.0\ORADATA\TEST\EXAMPLE01.DBF
   C:\TBS1.DBF
   C:\TBS2.DBF
   C:\TBS3.DBF
   C:\TBS4.DBF
   C:\TBS5.DBF
   C:\TBS6.DBF

   NAME
   --------------------------------------------------------------------------------
   C:\TBS7.DBF
   C:\TBS8.DBF
   C:\TEST1.DBF
   C:\WINDOWS\SYSTEM32\UNNAMED00015

   15 rows selected.

   SQL> alter database create datafile 'C:\WINDOWS\SYSTEM32\UNNAMED00015' as 'c:\test2.db

   Database altered.

   SQL> recover database;
   Media recovery complete.
   SQL> alter database open;

   Database altered.

   SQL> select * from testlost2;

   no rows selected

   SQL> select * from testlost;

   A
   ----------
   2
   1

   在上面这个试验里面我们可能还会有一个疑问,既然控制文件里面记载的是unamedxxxxxxx的文件名,数据字典的里面的filename也是 unamedxxxxxxx(实际上数据字典里面不保存文件名,所有的文件名都只包含在控制文件当中),日志里也是没有记载文件名,那么当我们 recover database的时候为什么会出现

   ORA-00283: recovery session canceled due to errors

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

   ORA-01110: data file 15: 'C:\TEST2.DBF'

   ORA-01112: media recovery not started

   Oracle(大型网站数据库平台)他是从哪里得到这个unamedxxxxxxx文件的真实文件是'C:\TEST2.DBF'呢?

   先来看看日志里记载了什么东西

   用logminer查一下

   EXECUTE DBMS_LOGMNR_D.BUILD(dictionary_filename => 'l_dictionary.ora',dictionary_location => '/disk1/oradata/');

   EXECUTE DBMS_LOGMNR.ADD_LOGFILE( LogFileName => '/disk1/oradata/redo04.log',Options => dbms_logmnr.NEW);

   EXECUTE DBMS_LOGMNR.START_LOGMNR(DictFileName =>'/disk1/oradata/l_dictionary.ora');

   SELECT sql_redo FROM V$LOGMNR_CONTENTS where upper(sql_redo) like '%TEST%';

   SQL_REDO
   --------------------------------------------------------------------------------
   create tablespace test3
   datafile  'c:\test3.dbf' size 1m reuse
   extent management local
   segment space management auto
   uniform size 40k;
   只发现了这句ddl.并没有往数据字典里插入文件名。

   再来看create tablespace的trace文件
   ……
   insert into ts$ (ts#,name,online$,contents$,undofile#,undoblock#,blocksize,
   dflmaxext,dflinit,dflincr,dflextpct,dflminext,dflminlen,inc#,owner#,scnwrp,
   scnbas,pitrscnwrp,pitrscnbas,dflogging, affstrength,bitmapped,plugged,
   directallowed,flags,spare1,spare2)
   values
   (:1,:2,:3,:4,:5,:6,:7,:8,:9,:10,:11,:12,:13,1,:14,:15,:16,:17,:18,:19,0,:20,:21,1,:22,:23,:24)

   insert into file$ (file#,blocks,ts#,status$,relfile#,maxextend,inc,crscnwrp,crscnbas,spare1)
   values
   (:1,:2,DECODE(:3,-1,NULL,:3),:4, DECODE(:5,0,NULL,:5),:6,:7,:8,:9,DECODE(:10,0,NULL,:10))
   ……

   这两条语句说明了Oracle(大型网站数据库平台)在ts$,file$分别插入了新创建表空间和文件的信息,但并不包括文件名。

   看起来好像这个文件名是无出处的,即不直接取至日志,也不存在于数据字典,控制文件中则存在着错误的文件名,那他究竟从哪里来呢?

   看下控制文件中包含这个文件的dump 信息

   DATA FILE #15:
   (name #4) C:\WINDOWS\SYSTEM32\UNNAMED00015
   creation size=0 block size=8192 status=0x2 head=4 tail=4 dup=1
   tablespace 16, index=15 krfil=15 prev_file=0
   unrecoverable scn: 0x0000.00000000 01/01/1988 00:00:00
   Checkpoint cnt:135 scn: 0x0000.001d4e00 12/02/2004 19:32:57
   Stop scn: 0x0000.001d4e00 12/02/2004 19:36:28
   Creation Checkpointed at scn:  0x0000.0012445c 08/25/2004 15:37:56

   从这里我们知道test2.dbf是16号文件,这样的话Oracle(大型网站数据库平台)可以从file$表中找到这个文件对应的ts#,再从ts$中找到表空间的名字,然后回到日志,对应日志中的ddl即可以得到错误的文件是c:\test2.dbf,答案出来了!Oracle(大型网站数据库平台)根据这么一套流程找到了正确的文件名并提示给用户,这样的话我们只需要很方便的create datafile reuse一下就可以继续恢复进程了。

上一页  1 2 3 4 5 6 7 8  下一页

Tags:Sql 数据库 MDF

编辑录入:coldstar [复制链接] [打 印]
赞助商链接