如何利用DBMS_LOGMNR包挖掘在线日志

今天实验内容是日志挖掘工具LOGMNR的使用，我的测试版本是10.2.0.1默认就自带了，无需另外安装。如果未安装过提示无法使用这个dbms包，则可以用SYSDBA登陆，然后依次执行：@$ORACLE_HOME\rdbms\admin\dbmslm.sql;@$ORACLE_HOME\rdbms\admin\dbmslmd.sql;第一

今天实验内容是日志挖掘工具LOGMNR的使用，我的测试版本是10.2.0.1默认就自带了，无需另外安装。 如果未安装过提示无法使用这个dbms包，则可以用SYSDBA登陆，然后依次执行： @$ORACLE_HOME\rdbms\admin\dbmslm.sql; @$ORACLE_HOME\rdbms\admin\dbmslmd.sql; 第一个脚本用来创建DBMS_LOGMNR包，该包用来分析日志文件。 第二个脚本用来创建DBMS_LOGMNR_D包，该包用来创建数据字典文件。
--开始实验
[oracle@ora10g ~]$ sqlplus / as sysdba
SQL*Plus: Release 10.2.0.1.0 - Production on 26 09:33:33 2014
Copyright (c) 1982, 2005, Oracle. All rights reserved.

Connected to: Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Production With the Partitioning, OLAP and Data Mining options
SQL> set line 130 SQL> select * from v$log;
GROUP# THREAD# SEQUENCE# BYTES MEMBERS ARC STATUS FIRST_CHANGE# FIRST_TIME ---------- ---------- ---------- ---------- ---------- --- ---------------- ------------- ---------- 1 1 8 52428800 1 YES INACTIVE 1402743 2014-12-24 2 1 9 52428800 1 NO CURRENT 1402823 2014-12-24 3 1 7 52428800 1 YES INACTIVE 1401824 2014-12-24
SQL> col member for a45 SQL> select group#,member from v$logfile;
GROUP# MEMBER ---------- --------------------------------------------- 3 /u01/app/oracle/oradata/ora10g/redo03.log 2 /u01/app/oracle/oradata/ora10g/redo02.log 1 /u01/app/oracle/oradata/ora10g/redo01.log
--启用日志挖掘
SQL> exec dbms_logmnr.add_logfile('/u01/app/oracle/oradata/ora10g/redo02.log');
PL/SQL procedure successfully completed.
SQL> exec dbms_logmnr.start_logmnr(options=>dbms_logmnr.dict_from_online_catalog);
PL/SQL procedure successfully completed.
来看一下LOGMNR工具用到的相关视图: SQL> set pages 100 SQL> col comments for a40 SQL> select * from dict t where t.table_name like '%LOGMNR%';
TABLE_NAME COMMENTS ------------------------------ ---------------------------------------- DBA_LOGMNR_LOG DBA_LOGMNR_SESSION DBA_LOGMNR_PURGED_LOG V$LOGMNR_CONTENTS Synonym for V_$LOGMNR_CONTENTS V$LOGMNR_PARAMETERS Synonym for V_$LOGMNR_PARAMETERS V$LOGMNR_DICTIONARY Synonym for V_$LOGMNR_DICTIONARY V$LOGMNR_LOGS Synonym for V_$LOGMNR_LOGS V$LOGMNR_STATS Synonym for V_$LOGMNR_STATS V$LOGMNR_DICTIONARY_LOAD Synonym for V_$LOGMNR_DICTIONARY_LOAD GV$LOGMNR_CONTENTS Synonym for GV_$LOGMNR_CONTENTS GV$LOGMNR_PARAMETERS Synonym for GV_$LOGMNR_PARAMETERS GV$LOGMNR_DICTIONARY Synonym for GV_$LOGMNR_DICTIONARY GV$LOGMNR_LOGS Synonym for GV_$LOGMNR_LOGS V$LOGMNR_LOGFILE Synonym for V_$LOGMNR_LOGFILE V$LOGMNR_PROCESS Synonym for V_$LOGMNR_PROCESS V$LOGMNR_LATCH Synonym for V_$LOGMNR_LATCH V$LOGMNR_TRANSACTION Synonym for V_$LOGMNR_TRANSACTION V$LOGMNR_REGION Synonym for V_$LOGMNR_REGION V$LOGMNR_CALLBACK Synonym for V_$LOGMNR_CALLBACK V$LOGMNR_SESSION Synonym for V_$LOGMNR_SESSION GV$LOGMNR_LOGFILE Synonym for GV_$LOGMNR_LOGFILE GV$LOGMNR_PROCESS Synonym for GV_$LOGMNR_PROCESS GV$LOGMNR_LATCH Synonym for GV_$LOGMNR_LATCH GV$LOGMNR_TRANSACTION Synonym for GV_$LOGMNR_TRANSACTION GV$LOGMNR_REGION Synonym for GV_$LOGMNR_REGION GV$LOGMNR_CALLBACK Synonym for GV_$LOGMNR_CALLBACK GV$LOGMNR_SESSION Synonym for GV_$LOGMNR_SESSION GV$LOGMNR_STATS Synonym for GV_$LOGMNR_STATS GV$LOGMNR_DICTIONARY_LOAD Synonym for GV_$LOGMNR_DICTIONARY_LOAD
29 rows selected.
这里主要用到的是v$logmnr_contents这个视图，里面存放里挖掘日志获得的内容，来看一下表结构:
SQL> desc v$logmnr_contents; Name Null? Type ----------------------------------------------------------------------- -------- ------------------------------------------------ SCN NUMBER CSCN NUMBER TIMESTAMP DATE COMMIT_TIMESTAMP DATE THREAD# NUMBER LOG_ID NUMBER XIDUSN NUMBER XIDSLT NUMBER XIDSQN NUMBER PXIDUSN NUMBER PXIDSLT NUMBER PXIDSQN NUMBER RBASQN NUMBER RBABLK NUMBER RBABYTE NUMBER UBAFIL NUMBER UBABLK NUMBER UBAREC NUMBER UBASQN NUMBER ABS_FILE# NUMBER REL_FILE# NUMBER DATA_BLK# NUMBER DATA_OBJ# NUMBER DATA_OBJD# NUMBER SEG_OWNER VARCHAR2(32) SEG_NAME VARCHAR2(256) TABLE_NAME VARCHAR2(32) SEG_TYPE NUMBER SEG_TYPE_NAME VARCHAR2(32) TABLE_SPACE VARCHAR2(32) ROW_ID VARCHAR2(18) SESSION# NUMBER SERIAL# NUMBER USERNAME VARCHAR2(30) SESSION_INFO VARCHAR2(4000) TX_NAME VARCHAR2(256) ROLLBACK NUMBER OPERATION VARCHAR2(32) OPERATION_CODE NUMBER SQL_REDO VARCHAR2(4000) SQL_UNDO VARCHAR2(4000) RS_ID VARCHAR2(32) SEQUENCE# NUMBER SSN NUMBER CSF NUMBER INFO VARCHAR2(32) STATUS NUMBER REDO_VALUE NUMBER UNDO_VALUE NUMBER SQL_COLUMN_TYPE VARCHAR2(30) SQL_COLUMN_NAME VARCHAR2(30) REDO_LENGTH NUMBER REDO_OFFSET NUMBER UNDO_LENGTH NUMBER UNDO_OFFSET NUMBER DATA_OBJV# NUMBER SAFE_RESUME_SCN NUMBER XID RAW(8) PXID RAW(8) AUDIT_SESSIONID NUMBER
--通过抓取关键字"create"，来获得刚才的建表语句
SQL> select scn,timestamp,log_id,seg_owner,seg_type,table_space,data_blk#,data_obj#,data_objd#,session#,serial#,username,session_info,sql_redo,sql_undo 2 from v$logmnr_contents t 3 where t.sql_redo like 'create%';

SCN TIMESTAMP LOG_ID SEG_OWNER SEG_TYPE TABLE_SPACE DATA_BLK# ---------- ---------- ---------- -------------------------------- ---------- -------------------------------- ---------- DATA_OBJ# DATA_OBJD# SESSION# SERIAL# USERNAME ---------- ---------- ---------- ---------- ------------------------------ SESSION_INFO ---------------------------------------------------------------------------------------------------------------------------------- SQL_REDO ---------------------------------------------------------------------------------------------------------------------------------- SQL_UNDO ---------------------------------------------------------------------------------------------------------------------------------- 1404138 2014-12-26 9 ZLM 2 0 53863 0 0 0
create table zlm.t1 as select * from dba_objects;
可以看到，刚才的DDL建表语句已经被记录到v$logmnr_contents表中了
--结束日志挖掘
SQL> exec dbms_logmnr.end_logmnr;
PL/SQL procedure successfully completed.
一旦结束日志挖掘后，就无法再对v$logmgr_contents视图进行查询，会报错：
SQL> select scn,timestamp,log_id,seg_owner,seg_type,table_space,data_blk#,data_obj#,data_objd#,session#,serial#,username,session_info,sql_redo,sql_undo 2 from v$logmnr_contents t 3 where t.sql_redo like 'create%'; from v$logmnr_contents t * ERROR at line 2: ORA-01306: dbms_logmnr.start_logmnr() must be invoked before selecting from v$logmnr_contents
并且要重新指定挖掘的日志对象，否则依然报错：
SQL> exec dbms_logmnr.start_logmnr(options=>dbms_logmnr.dict_from_online_catalog); BEGIN dbms_logmnr.start_logmnr(options=>dbms_logmnr.dict_from_online_catalog); END;
【本文来自鸿网互联 (http://www.68idc.cn)】 * ERROR at line 1: ORA-01292: no log file has been specified for the current LogMiner session ORA-06512: at "SYS.DBMS_LOGMNR", line 58 ORA-06512: at line 1
SQL> create table zlm.t2 as select * from zlm.t1; create table zlm.t2 as select * from zlm.t1 * ERROR at line 1: ORA-01652: unable to extend temp segment by 128 in tablespace WEB
由于之前zlm用户给的默认表空间WEB空间比较小，只有10M，因此报空间不足了
SQL> select d.file#,d.ts#,d.bytes/1024/1024 as "Size(M)",t.name from v$datafile d,v$tablespace t where d.ts#=t.ts#;
FILE# TS# Size(M) NAME ---------- ---------- ---------- --------------------------------------------- 1 0 560 SYSTEM 2 1 165 UNDOTBS1 3 2 270 SYSAUX 4 4 40 USERS 5 6 100 EXAMPLE 6 7 50 ZLM 7 18 20 INDX2 8 19 10 WEB
8 rows selected.
--修改zlm缺省表空间为zlm
SQL> alter user zlm default tablespace zlm;
User altered.
SQL> create table zlm.t2 as select * from zlm.t1;
Table created.
SQL> select * from v$log;
GROUP# THREAD# SEQUENCE# BYTES MEMBERS ARC STATUS FIRST_CHANGE# FIRST_TIME ---------- ---------- ---------- ---------- ---------- --- ---------------- ------------- ---------- 1 1 11 52428800 1 NO CURRENT 1405472 2014-12-26 2 1 9 52428800 1 YES INACTIVE 1402823 2014-12-24 3 1 10 52428800 1 YES INACTIVE 1405357 2014-12-26
由于之前创建t1表的操作会产生一定的redo,并且是一个DDL操作，会自动commit，当前日志已经切换到group 1了 SQL> select scn,timestamp,log_id,seg_owner,seg_type,table_space,data_blk#,data_obj#,data_objd#,session#,serial#,username,session_info,sql_redo,sql_undo 2 from v$logmnr_contents t 3 where t.sql_redo like 'create%';
SCN TIMESTAMP LOG_ID SEG_OWNER SEG_TYPE TABLE_SPACE DATA_BLK# ---------- ---------- ---------- -------------------------------- ---------- -------------------------------- ---------- DATA_OBJ# DATA_OBJD# SESSION# SERIAL# USERNAME ---------- ---------- ---------- ---------- ------------------------------ SESSION_INFO ---------------------------------------------------------------------------------------------------------------------------------- SQL_REDO ---------------------------------------------------------------------------------------------------------------------------------- SQL_UNDO ---------------------------------------------------------------------------------------------------------------------------------- 1404138 2014-12-26 9 ZLM 2 0 53863 0 0 0
create table zlm.t1 as select * from dba_objects;
既然当前日志是group 1的日志文件，而之前挖掘对象是group 2的日志文件，所以查询v$logmgr_contetns视图发现，第2条建表语句并没有被记录下来
--重新对group 1的redo01.log日志进行挖掘
SQL> exec dbms_logmnr.add_logfile('/u01/app/oracle/oradata/ora10g/redo01.log');
PL/SQL procedure successfully completed.
SQL> exec dbms_logmnr.start_logmnr(options=>dbms_logmnr.dict_from_online_catalog);
PL/SQL procedure successfully completed.
--查看LOGMNR内容表
SQL> select scn,timestamp,log_id,seg_owner,seg_type,table_space,data_blk#,data_obj#,data_objd#,session#,serial#,username,session_info,sql_redo,sql_undo 2 from v$logmnr_contents t 3 where t.sql_redo like 'create%';
SCN TIMESTAMP LOG_ID SEG_OWNER SEG_TYPE TABLE_SPACE DATA_BLK# ---------- ---------- ---------- -------------------------------- ---------- -------------------------------- ---------- DATA_OBJ# DATA_OBJD# SESSION# SERIAL# USERNAME ---------- ---------- ---------- ---------- ------------------------------ SESSION_INFO ---------------------------------------------------------------------------------------------------------------------------------- SQL_REDO ---------------------------------------------------------------------------------------------------------------------------------- SQL_UNDO ---------------------------------------------------------------------------------------------------------------------------------- 1404138 2014-12-26 9 ZLM 2 0 53863 0 0 0
create table zlm.t1 as select * from dba_objects;

1406186 2014-12-26 11 ZLM 2 0 53864 0 0 0
create table zlm.t2 as select * from zlm.t1;

1406401 2014-12-26 11 ZLM 2 0 53865 0 0 0
create table zlm.t2 as select * from zlm.t1;
对redo01.log日志挖掘后，新的两条创建t2表的语句就被写入v$logmnr_contents中了，包括之前由于空间不足创建失败的那条语句，也做了记录，他们各自都有自己的scn号，不会互相冲突。注意，挖掘日志完毕以后，记得用exec dbms_logmnr.end_logmnr;关闭，否则会消耗数据库资源。
总结：
通过使用dbms_logmnr包，可以对在线日志进行挖掘，来获对数据库操作相关语句，对于审计也有一定的作用。高级用法还有很多，包括对归档日志的挖掘，对其他内容的挖掘等，这里只是做了一个简单的测试，觉得这个工具还是挺好用的。