如何在Oracle數(shù)據(jù)庫中找出損壞索引？

在Oracle數(shù)據(jù)庫中如何找出損壞索引呢？ 下面我們?nèi)藶闃?gòu)造一個案例，將索引塊損壞。如下案例所示：

SQL> create tablespace test_data

2 datafile '/u01/app/oracle/oradata/gsp/test_data_01.dbf'

3 size 200M autoextend off

4 logging

5 segment space management auto

6 extent management local;

Tablespace created.

SQL> create tablespace test_index

2 datafile '/u01/app/oracle/oradata/gsp/test_idx_01.dbf'

3 size 200M autoextend off

4 logging

5 segment space management auto

6 extent management local;

Tablespace created.

SQL> create user kerry

2 identified by 123456

3 default tablespace test_data;

User created.

SQL> grant connect to kerry;

SQL> grant resource to kerry;

上述腳本是創(chuàng)建表空間，創(chuàng)建用戶kerry并授權(quán)，然后使用kerry賬號登錄數(shù)據(jù)庫，構(gòu)造測試數(shù)據(jù)，在TEST表上創(chuàng)建索引IX_TEST

SQL> show user;

USER is "KERRY"

SQL>

SQL> CREATE TABLE TEST(ID NUMBER(10), NAME VARCHAR2(64));

Table created.

SQL> DECLARE I NUMBER;

2 BEGIN

3 FOR I IN 1..1000 LOOP

4 INSERT INTO TEST VALUES(I, LPAD('T', 60));

5 END LOOP;

6 COMMIT;

7 END;

8 /

PL/SQL procedure successfully completed.

SQL> CREATE INDEX IX_TEST ON KERRY.TEST(NAME) TABLESPACE TEST_INDEX;

Index created.

然后使用下面腳本找到索引段數(shù)據(jù)庫文件ID，以及索引段的第一個塊的塊號。

SQL> show user;

USER is "SYS"

SQL> col segment_name for a32;

SQL> col header_file for 9999;

SQL> col header_block for 9999;

SQL> select segment_name

2 ,header_file

3 ,header_block

4 ,blocks

5 from dba_segments ds

6 where ds.owner='KERRY' and ds.segment_name='IX_TEST';

SEGMENT_NAME HEADER_FILE HEADER_BLOCK BLOCKS

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

IX_TEST 8 130 16

SQL>

構(gòu)造壞塊的方法有不少（例如BBED等），這里我們使用RMAN下面的命令clear,可以標(biāo)記數(shù)據(jù)塊為corrupt,標(biāo)記數(shù)據(jù)文件8中130號數(shù)據(jù)塊為壞塊。

[oracle@DB-Server ~]$ rman target /

Recovery Manager: Release 11.2.0.1.0 - Production on Thu Sep 13 17:41:05 2018

Copyright (c) 1982, 2009, Oracle and/or its affiliates. All rights reserved.

connected to target database: GSP (DBID=644393201)

RMAN> recover datafile 8 block 130 clear;

Starting recover at 13-SEP-18

using target database control file instead of recovery catalog

allocated channel: ORA_DISK_1

channel ORA_DISK_1: SID=12 device type=DISK

Finished recover at 13-SEP-18

RMAN>

那么我們先來看看使用那些方法驗(yàn)證索引損壞了，測試驗(yàn)證一下看看是否可行。

1：使用ANALYZE分析驗(yàn)證索引結(jié)構(gòu)

[oracle@DB-Server ~]$ sqlplus / as sysdba

SQL*Plus: Release 11.2.0.1.0 Production on Thu Sep 13 17:42:03 2018

Copyright (c) 1982, 2009, Oracle. All rights reserved.

Connected to:

Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - 64bit Production

With the Partitioning, OLAP, Data Mining and Real Application Testing options

SQL> analyze index kerry.ix_test validate structure;

analyze index kerry.ix_test validate structure

*

ERROR at line 1:

ORA-01578: ORACLE data block corrupted (file # 8, block # 130)

ORA-01110: data file 8: '/u01/app/oracle/oradata/gsp/test_idx_01.dbf'

如上截圖所示，如果索引損壞（Corrupt Index），那么使用analyze index validate structure就會報(bào)錯。要檢查整個數(shù)據(jù)庫所有的損壞索引（Corrupt Indexes）的話，就可以借助下面腳本：

spool analy_index.sql

SET PAGESIZE 50000;

SELECT

'ANALYZE INDEX ' || OWNER || '.' || INDEX_NAME|| ' VALIDATE STRUCTURE;' FROM DBA_INDEXES;

spool off;

@analy_index.sql

2：使用系統(tǒng)視圖v$database_block_corruption查看損壞索引

如下所示，我們使用這個腳本來查看出現(xiàn)壞塊的索引，發(fā)現(xiàn)這個腳本無法找出壞塊索引。

set pagesize 50 linesize 170

col segment_name format a30

col partition_name format a30

SELECT DISTINCT file#,

segment_name,

segment_type,

tablespace_name,

partition_name

FROM dba_extents a,

v$database_block_corruption b

WHERE a.file_id = b.file#

AND a.block_id <= b.block#?

AND a.block_id + a.blocks >= b.block#;

原因分析如下，視圖v$database_block_corruption中有壞塊記錄，但是我們將索引段的第一個塊標(biāo)記為壞塊后，在dba_extents中沒有該索引段的記錄了。所以這種情況下的索引損壞，這個SQL語句根本無法找出壞塊索引。

SQL> SELECT file_id,

2 segment_name,

3 segment_type

4 FROM dba_extents

5 WHERE file_id = 8 ;

no rows selected

SQL> SELECT file_id,

2 segment_name,

3 segment_type

4 FROM dba_extents

5 WHERE owner = 'KERRY';

FILE_ID SEGMENT_NAME SEGMENT_TYPE

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

7 TEST TABLE

7 TEST TABLE

SQL>

因?yàn)橐粋€段的第一個區(qū)的第一個塊是FIRST LEVEL BITMAP BLOCK，第二個塊是SECOND LEVEL BITMAP BLOCK，這兩個塊是用來管理free block的，第三個塊是PAGETABLE SEGMENT HEADER，這個塊才是segment里的HEADER_BLOCK，再后面的塊就是用來記錄數(shù)據(jù)的。關(guān)于這些知識，可以參考我博客ORACLE關(guān)于段的HEADER_BLOCK的一點(diǎn)淺析。而我們最上面的例子，是將第一個塊構(gòu)造為壞塊，所以導(dǎo)致上面SQL無法查出。 我們重新構(gòu)造案例，如我們將索引段的數(shù)據(jù)塊構(gòu)造為壞塊，例如下面，將塊號148人為構(gòu)造壞塊。那么此時這個腳本就能找出壞塊索引了。所以綜上述實(shí)驗(yàn)可以看出，這個腳本查找壞塊索引是有條件的，要看索引段損壞的塊是什么類型

SQL> SELECT FILE_ID,

2 BLOCK_ID,

3 BLOCKS

FROM DBA_EXTENTS

4 5 WHERE OWNER ='&OWNER'

6 AND SEGMENT_NAME = '&TABLE_NAME';

Enter value for owner: KERRY

old 5: WHERE OWNER ='&OWNER'

new 5: WHERE OWNER ='KERRY'

Enter value for table_name: IX_TEST

old 6: AND SEGMENT_NAME = '&TABLE_NAME'

new 6: AND SEGMENT_NAME = 'IX_TEST'

FILE_ID BLOCK_ID BLOCKS

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

8 144 8

8 152 8

SQL> SELECT HEADER_FILE

2 , HEADER_BLOCK

3 , BYTES

4 , BLOCKS

5 , EXTENTS

FROM DBA_SEGMENTS

6 7 WHERE OWNER='&OWNER' AND SEGMENT_NAME='&SEGMENT_NAME';

Enter value for owner: KERRY

Enter value for segment_name: IX_TEST

old 7: WHERE OWNER='&OWNER' AND SEGMENT_NAME='&SEGMENT_NAME'

new 7: WHERE OWNER='KERRY' AND SEGMENT_NAME='IX_TEST'

HEADER_FILE HEADER_BLOCK BYTES BLOCKS EXTENTS

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

8 146 131072 16 2

SQL>

RMAN> recover datafile 8 block 148 clear;