利用RMAN修复数据文件中的坏块
 
1、创建测试表：
SQL> create table tab_dbv tablespace ghost as select * from dba_objects;
表已创建。
 
SQL> insert into tab_dbv select * from tab_dbv;
已创建50363行。
 
SQL> insert into tab_dbv select * from tab_dbv;
已创建100726行。
 
SQL> commit;
提交完成。
 
2、查找tab_dbv的块头：
SQL> select segment_name,header_block from dba_segments where segment_name='TAB_DBV';
 
SEGMENT_NAME              HEADER_BLOCK
------------------------- ------------
TAB_DBV                             35
 
SQL> select to_char(35*8*1024,'xxxxxxxxxxxxxxxxxxxxxx') from dual;
 
TO_CHAR(35*8*1024,'XXXX
-----------------------
                  46000
 
SQL> select to_char(36*8*1024,'xxxxxxxxxxxxxxxxxxxxxx') from dual;
 
TO_CHAR(36*8*1024,'XXXX
-----------------------
                  48000
 
我们使用UE修改46000H到48000H之间的数据。
 
启动数据库，查询：
SQL> select count(*) from tab_dbv;
select count(*) from tab_dbv
                     *
ERROR at line 1:
ORA-01578: ORACLE data block corrupted (file # 6, block # 560)
ORA-01110: data file 6: 'E:\ORACLE\PRODUCT\10.2.0\ORADATA\HUNT\GHOST_001.DBF'
 
使用DBV对数据文件进行检查：
C:\>dbv file='E:\oracle\product\10.2.0\oradata\hunt\GHOST_001.DBF'
......
Page 560 is marked corrupt
Corrupt block relative dba: 0x01800230 (file 6, block 560)  ---第一个坏块位置
……
Page 576 is marked corrupt
Corrupt block relative dba: 0x01800240 (file 6, block 576)  ---第二个坏块位置
 
DBVERIFY - Verification complete
 
Total Pages Examined         : 14080
Total Pages Processed (Data) : 2782
Total Pages Failing   (Data) : 0
Total Pages Processed (Index): 88
Total Pages Failing   (Index): 0
Total Pages Processed (Other): 80
Total Pages Processed (Seg)  : 0
Total Pages Failing   (Seg)  : 0
Total Pages Empty            : 11128
Total Pages Marked Corrupt   : 2               -----有两个块损坏。
Total Pages Influx           : 0
Highest block SCN            : 1255980 (0.1255980)
 
2、检查坏块：
RMAN> backup validate database;
SQL>select * from v$database_block_corruption;
  
FILE#
BLOCK#
BLOCKS
CORRUPTION_CHANGE#
CORRUPTION_TYPE
1
6
576
1
0
CORRUPT
2
6
560
1
0
CORRUPT
与DBV报出的结果一致。
 
3、修复坏块：
RMAN> blockrecover datafile 6 block 560;
---此命令是10G才有的，9I不支持。
RMAN> backup validate datafile 6;
---重新检验。
SQL>select * from v$database_block_corruption;
  
FILE#
BLOCK#
BLOCKS
CORRUPTION_CHANGE#
CORRUPTION_TYPE
1
6
576
1
0
CORRUPT
SQL> select count(*) from tab_dbv;
select count(*) from tab_dbv
                     *
ERROR at line 1:
ORA-01578: ORACLE data block corrupted (file # 6, block # 576) 
---此时报576损坏，说明560已修复
ORA-01110: data file 6: 'E:\ORACLE\PRODUCT\10.2.0\ORADATA\HUNT\GHOST_001.DBF'
说明：
存在坏块时RMAN备份无法进行，如：
RMAN> backup datafile 6;
 
Starting backup at 26-MAY-12
using channel ORA_DISK_1
channel ORA_DISK_1: starting full datafile backupset
channel ORA_DISK_1: specifying datafile(s) in backupset
input datafile fno=00006 name=E:\ORACLE\PRODUCT\10.2.0\ORADATA\HUNT\GHOST_001.DBF
channel ORA_DISK_1: starting piece 1 at 26-MAY-12
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03009: failure of backup command on ORA_DISK_1 channel at 05/26/2012 08:54:40
ORA-19566: exceeded limit of 0 corrupt blocks for file E:\ORACLE\PRODUCT\10.2.0\ORADATA\HUNT\GHOST_001.DBF
 
我们可以修改corrupt的允许数量（默认为0）对数据文件进行备份。在执行RUN脚本时指定maxcorrupt即可：
RUN {
Set maxcorrupt for datafile 6 to 1;
Backup datafile 6;
…..
}
在命令行方式下执行会报错：
RMAN> Set maxcorrupt for datafile 6 to 1;
 
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03031: this option of set command needs to be used inside a run block
 
按同样方法修复(file # 6, block # 576)：
RMAN> blockrecover datafile 6 block 576;
RMAN> backup validate datafile 6;
SQL>select * from v$database_block_corruption;
no rows selected
查询数据：
SQL> select count(*) from tab_dbv;
 
  COUNT(*)
----------
201452
 
修复成功！