不完全恢复的基本类型：
1）基于时间点 （until time): 使整个数据库恢复到过去的一个时间点前
2）基于scn （until change）： 使整个数据库恢复到过去的某个SCN前
3）基于cancel (until cancel)： 使整个数据库恢复到归档日志或当前日志的断点前
不完全恢复（Incomplete recover） 适用环境：
1）在过去的某个时间点重要的数据被破坏。
2）在做完全恢复时，丢失了归档日志或当前online redo log
3）当误删除了表空间时（有控制文件备份）
4）丢失了所有的控制文件，使用备份的控制文件恢复时 （条件满足时可以完全恢复）
传统的不完全恢复的操作步骤：
1）先通过logmnr 找到误操作的时间点
2）对现在的database做新全备
3）还原该时间点前所有的datafile
4）在mount状态下，对database做recover，恢复到误操作的时间点
5）将恢复出来的table做逻辑备份（exp）
6）再将全备还原
7）将导出的表导入database（imp）
实验 1：恢复过去某个时间点误操作的table （基于时间点的不完全恢复） 前提：有冷备份，日志，归档齐全
1）准备实验数据
SQL> select * from andy;
ID
----------
4
SQL> insert into andy values(5);
1 row created.
SQL> commit;
Commit complete.
SQL> drop table andy purge;
Table dropped.
2）查看日志，归档环境
SQL> set linesize 400
SQL> select * from v$log;
GROUP# THREAD# SEQUENCE# BYTES BLOCKSIZE MEMBERS ARC STATUS FIRST_CHANGE#
---------- ---------- ---------- ---------- ---------- ---------- --- ---------------- -------------
1 1 61 52428800 512 1 NO CURRENT 1991534
2 1 59 52428800 512 1 YES INACTIVE 1959769
3 1 60 52428800 512 1 YES INACTIVE 1966481
SQL> alter system switch logfile;
System altered.
SQL> select name from v$archived_log;
NAME
----------------------------------------------------------------------------------------------------------
/home/oracle/archivelog/ORCL/archivelog/2014_12_01/o1_mf_1_47_b7rwlclg_.arc
/home/oracle/app/flash_recovery_area/ORCL/archivelog/2014_12_01/o1_mf_1_48_b7ryjgng_.arc
/home/oracle/app/flash_recovery_area/ORCL/archivelog/2014_12_01/o1_mf_1_49_b7rykz3l_.arc
/home/oracle/app/flash_recovery_area/ORCL/archivelog/2014_12_01/o1_mf_1_50_b7ryn3fl_.arc
/home/oracle/archivelog/1_51_860522448.dbf
/home/oracle/archivelog/1_52_860522448.dbf
/home/oracle/archivelog/1_53_860522448.dbf
/home/oracle/archivelog/1_54_860522448.dbf
/home/oracle/archivelog/1_55_860522448.dbf
NAME
----------------------------------------------------------------------------------------------------------
/home/oracle/archivelog/1_56_860522448.dbf
/home/oracle/archivelog/1_57_860522448.dbf
/home/oracle/archivelog/1_58_860522448.dbf
/home/oracle/archivelog/1_59_860522448.dbf
/home/oracle/archivelog/1_60_860522448.dbf
/home/oracle/archivelog/1_61_860522448.dbf
28 rows selected.
3）logmnr日志挖掘，找出purge时间点。
--日志挖掘至少要提前开SUPPLEME，如果没开，信息会报错
SQL> select supplemental_log_data_min,supplemental_log_data_pk,supplemental_log_data_ui from v$database;
SUPPLEME SUP SUP
-------- --- ---
YES NO NO
SQL> col member for a50;
SQL> select * from v$logfile;
GROUP# STATUS TYPE MEMBER IS_
---------- ------- ------- -------------------------------------------------- ---
3 ONLINE /home/oracle/app/oradata/orcl/redo03.log NO
2 ONLINE /home/oracle/app/oradata/orcl/redo02.log NO
1 ONLINE /home/oracle/app/oradata/orcl/redo01.log NO
SQL> select group#,status from v$log;
GROUP# STATUS
---------- ----------------
1 ACTIVE
2 CURRENT
3 INACTIVE
SQL> select name from v$archived_log where name is not null order by 1;
NAME
-----------------------------------------------------------------------------------------
/home/oracle/app/flash_recovery_area/ORCL/archivelog/2014_12_01/o1_mf_1_48_b7ryjgng_.arc
/home/oracle/app/flash_recovery_area/ORCL/archivelog/2014_12_01/o1_mf_1_49_b7rykz3l_.arc
/home/oracle/app/flash_recovery_area/ORCL/archivelog/2014_12_01/o1_mf_1_50_b7ryn3fl_.arc
/home/oracle/archivelog/1_51_860522448.dbf
/home/oracle/archivelog/1_52_860522448.dbf
/home/oracle/archivelog/1_53_860522448.dbf
/home/oracle/archivelog/1_54_860522448.dbf
/home/oracle/archivelog/1_55_860522448.dbf
/home/oracle/archivelog/1_56_860522448.dbf
/home/oracle/archivelog/1_57_860522448.dbf
/home/oracle/archivelog/1_58_860522448.dbf
NAME
-----------------------------------------------------------------------------------------
/home/oracle/archivelog/1_59_860522448.dbf
/home/oracle/archivelog/1_60_860522448.dbf
/home/oracle/archivelog/1_61_860522448.dbf
/home/oracle/archivelog/ORCL/archivelog/2014_12_01/o1_mf_1_47_b7rwlclg_.arc
15 rows selected.
-- 利用redolog日志进行挖掘
SQL> EXECUTE DBMS_LOGMNR.ADD_LOGFILE(LogFileName=>'/home/oracle/app/oradata/orcl/redo01.log',Options=>dbms_logmnr.new);
PL/SQL procedure successfully completed.
SQL>EXECUTE DBMS_LOGMNR.ADD_LOGFIL (LogFileName=>'/home/oracle/app/oradata/orcl/redo02.log',Options=>dbms_logmnr.ADDFILE);
PL/SQL procedure successfully completed.
SQL>EXECUTE DBMS_LOGMNR.ADD_LOGFILE(LogFileName=>'/home/oracle/app/oradata/orcl/redo03.log',Options=>dbms_logmnr.ADDFILE);
PL/SQL procedure successfully completed.
SQL> execute DBMS_LOGMNR.START_LOGMNR(options=>dbms_logmnr.dict_from_online_catalog);
PL/SQL procedure successfully completed.
SQL> col username for a10;
SQL> col sql_redo for a45;
SQL> select username,scn,timestamp,sql_redo from v$logmnr_contents where seg_name='ANDY' order by scn;
USERNAME SCN TIMESTAMP SQL_REDO
---------- ---------- ------------------- ---------------------------------------------
ANDY 2000934 2014-12-11 09:47:15 drop table andy purge; //找到purge时间
--利用归档进行日志挖掘
SQL> show parameter utl
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
create_stored_outlines string
utl_file_dir string
SQL> alter system set utl_file_dir='/home/oracle/logmnr' scope=spfile;
System altered.
SQL> startup force;
SQL> show parameter utl;
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
create_stored_outlines string
utl_file_dir string /home/oracle/logmnr
SQL> execute dbms_logmnr_d.build('dict.ora','/home/oracle/logmnr',dbms_logmnr_d.store_in_flat_file);
PL/SQL procedure successfully completed.
SQL> execute dbms_logmnr.add_logfile(logfilename=>'/home/oracle/archivelog/1_61_860522448.dbf',options=>dbms_logmnr.new);
PL/SQL procedure successfully completed.
SQL> execute dbms_logmnr.start_logmnr(dictfilename=>'/home/oracle/logmnr/dict.ora',options=>dbms_logmnr.ddl_dict_tracking);
PL/SQL procedure successfully completed.
SQL> select username,scn,to_char(timestamp,'yyyy-mm-dd hh24:mi:ss'),sql_redo from v$logmnr_contents WHERE lower(sql_redo) like 'drop table%';
USERNAME SCN TO_CHAR(TIMESTAMP,' SQL_REDO
------------------------------ ---------- ------------------- --------------------------------------------------
ANDY 2000934 2014-12-11 09:47:15 drop table andy purge;
SQL> execute dbms_logmnr.end_logmnr;
PL/SQL procedure successfully completed.
4)关闭数据库，删除所有dbf，准备做不完全恢复
SQL> shutdown immdiate;
[oracle@11g logmnr]$ cd /home/oracle/app/oradata/orcl/
[oracle@11g orcl]$ rm -rf *.dbf
5)还原所有备份的数据文件
[oracle@11g orcl]$ cp /home/oracle/coldbak/*.dbf /home/oracle/app/oradata/orcl/
6)根据log miner提供的信息，做基于时间点的不完全恢复
SQL> recover database until time '2014-12-11 09:47:15';
ORA-00279: change 1968596 generated at 12/10/2014 06:26:35 needed for thread 1
ORA-00289: suggestion : /home/oracle/archivelog/1_60_860522448.dbf
ORA-00280: change 1968596 for thread 1 is in sequence #60
Specify log: {<RET>=suggested | filename | AUTO | CANCEL}
auto
ORA-00279: change 1991534 generated at 12/10/2014 07:25:45 needed for thread 1
ORA-00289: suggestion : /home/oracle/archivelog/1_61_860522448.dbf
ORA-00280: change 1991534 for thread 1 is in sequence #61
Log applied.
Media recovery complete.
说明：如果恢复过程中用到归档日志则输入 auto 。用到当前日志，则输入 filename 。
7)resetlogs方式打开数据库
SQL> alter database open resetlogs;
Database altered.
8)验证
SQL> select * from andy;
ID
----------
5
4
OK，转载请标明出处。
本文转自 张冲andy 博客园博客，原文链接：http://www.cnblogs.com/andy6/p/6274603.html ，如需转载请自行联系原作者