首页 > 代码库 > [转]开启闪回以及闪回的四种原理
[转]开启闪回以及闪回的四种原理
1、首先确认db_recovery_file_dest和db_recovery_file_dest_size 有值。
sys@TEST0910> show parameter recovery
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
db_recovery_file_dest string /u01/app/oracle/fast_recovery_ area
db_recovery_file_dest_size big integer 4122M
db_recovery_file_dest 放闪回日志,默认路径/u01/app/oracle/fast_recovery_ area
db_recovery_file_dest_size 存放闪回日志空间的大小,默认4122M
2、开启闪回
sys@TEST0910> shutdown immediate
Database closed.
Database dismounted.
ORACLE instance shut down.
sys@TEST0910> startup mount
ORACLE instance started.
Total System Global Area 2505338880 bytes
Fixed Size 2230952 bytes
Variable Size 587203928 bytes
Database Buffers 1895825408 bytes
Redo Buffers 20078592 bytes
Database mounted.
sys@TEST0910> alter database flashback on;
Database altered.
sys@TEST0910> alter database open;
Database altered.
3、查询确认是否开启闪回
sys@TEST0910> select name,flashback_on from v$database;
NAME FLASHBACK_ON
-------------------------------------------------- ------------------
TEST0910 YES
闪回四种不同原理
- flashback drop:利用recyelebin 回收站原理
- flashback table TEST10 to before drop [rename to test11];可以重命名
- flashback query ,依赖undo数据,先查询,在做insert
- insert into test10 select * from test10 as of timestamp to_timestamp(‘2013-09-23 11:52:06‘,‘yyyy-mm-dd hh24:mi:ss‘);
- flashback table 依赖undo块
- flashback table scott.test10 to scn 1952615
- insert into test10 select * from scott.test10 as of scn 1952615;
- flashback database DDL语句,依赖闪回区的闪回日志
- flashback database to timestamp to_timestamp(‘2013-09-23 15:09:52‘,‘yyyy-mm-dd hh24:mi:ss‘);时间减一秒。
- 闪回归档--create as
来源:http://blog.csdn.net/rlhua/article/details/11934831
flashback drop实验(sys用户的除外,它没有回收站)
1、查询原来表的数量
sys@TEST0910> select count(*) from scott.test10;
COUNT(*)
----------
14
2、误操作drop
sys@TEST0910> drop table scott.test10;
Table dropped.
3、登陆scoot用户,查询回收站
sys@TEST0910> conn scott/tiger
Connected.
scott@TEST0910> show recyclebin
ORIGINAL NAME RECYCLEBIN NAME OBJECT TYPE DROP TIME
---------------- ------------------------------ ------------ -------------------
TEST10 BIN$5w/YlwePKa7gQwEAAH+tZw==$0 TABLE 2013-09-23:11:40:56
4、查询回收站里的内容
scott@TEST0910> select * from "BIN$5w/YlwePKa7gQwEAAH+tZw==$0"; ------注意:此处是双引号
EMPNO ENAME JOB MGR HIREDATE SAL COMM DEPTNO
---------- ---------- --------- ---------- --------- ---------- ---------- ----------
7369 SMITH CLERK 7902 17-DEC-80 800 20
7499 ALLEN SALESMAN 7698 20-FEB-81 1600 300 30
7521 WARD SALESMAN 7698 22-FEB-81 1250 500 30
7566 JONES MANAGER 7839 02-APR-81 2975 20
7654 MARTIN SALESMAN 7698 28-SEP-81 1250 1400 30
7698 BLAKE MANAGER 7839 01-MAY-81 2850 30
7782 CLARK MANAGER 7839 09-JUN-81 2450 10
7788 SCOTT ANALYST 7566 19-APR-87 3000 20
7839 KING PRESIDENT 17-NOV-81 5000 10
7844 TURNER SALESMAN 7698 08-SEP-81 1500 0 30
7876 ADAMS CLERK 7788 23-MAY-87 1100 20
7900 JAMES CLERK 7698 03-DEC-81 950 30
7902 FORD ANALYST 7566 03-DEC-81 3000 20
7934 MILLER CLERK 7782 23-JAN-82 1300 10
14 rows selected.
5、两种闪回方式
使用insert into TEST10 select * from "BIN$5w/YlwePKa7gQwEAAH+tZw==$0"; 闪回
使用flashback table闪回
scott@TEST0910> flashback table TEST10 to before drop [rename to test11];可以重命名
Flashback complete.
scott@TEST0910> select count(*) from test10;
COUNT(*)
----------
14
flashback query实验(基于undo数据)
误操作DML语句,刚刚发生,undo没有被覆盖
1、查询原数据
11:52:06 scott@TEST0910> select count(*) from test10;
COUNT(*)
----------
14
11:52:06 这个时间点还有数据,稍后作为参考
2、误操作
11:52:19 scott@TEST0910> delete from test10;
14 rows deleted.
11:52:31 scott@TEST0910> commit;
Commit complete.
11:52:35 scott@TEST0910> select count(*) from test10;
COUNT(*)
----------
0
3、使用闪回查询
11:54:25 scott@TEST0910> select * from test10 as of timestamp to_timestamp(‘2013-09-23 11:52:06‘,‘yyyy-mm-dd hh24:mi:ss‘);
EMPNO ENAME JOB MGR HIREDATE SAL COMM DEPTNO
---------- ---------- --------- ---------- --------- ---------- ---------- ----------
7369 SMITH CLERK 7902 17-DEC-80 800 20
7499 ALLEN SALESMAN 7698 20-FEB-81 1600 300 30
7521 WARD SALESMAN 7698 22-FEB-81 1250 500 30
7566 JONES MANAGER 7839 02-APR-81 2975 20
7654 MARTIN SALESMAN 7698 28-SEP-81 1250 1400 30
7698 BLAKE MANAGER 7839 01-MAY-81 2850 30
7782 CLARK MANAGER 7839 09-JUN-81 2450 10
7788 SCOTT ANALYST 7566 19-APR-87 3000 20
7839 KING PRESIDENT 17-NOV-81 5000 10
7844 TURNER SALESMAN 7698 08-SEP-81 1500 0 30
7876 ADAMS CLERK 7788 23-MAY-87 1100 20
7900 JAMES CLERK 7698 03-DEC-81 950 30
7902 FORD ANALYST 7566 03-DEC-81 3000 20
7934 MILLER CLERK 7782 23-JAN-82 1300 10
14 rows selected.
4、插入闪回查询的结果
11:54:44 scott@TEST0910> insert into test10 select * from test10 as of timestamp to_timestamp(‘2013-09-23 11:52:06‘,‘yyyy-mm-dd hh24:mi:ss‘);
14 rows created.
11:55:02 scott@TEST0910> commit;
Commit complete.
11:55:30 scott@TEST0910> select count(*) from test10;
COUNT(*)
----------
14
flashback table实验(基于undo数据)
1、查询原数据
sys@TEST0910> select count(*) from test10;
COUNT(*)
----------
14
1 row selected.
2、查询数据库当前scn号,稍后作为恢复参考
sys@TEST0910> select checkpoint_change# from v$database;
CHECKPOINT_CHANGE#
------------------
1952615
1 row selected.
3、误操作
sys@TEST0910> delete from scott.test10;
14 rows deleted.
sys@TEST0910> commit;
Commit complete.
4、根据之前的scn号使用undo数据闪回查询
sys@TEST0910> select * from scott.test10 as of scn 1952615;
EMPNO ENAME JOB MGR HIREDATE SAL COMM DEPTNO
---------- ---------- --------- ---------- --------- ---------- ---------- ----------
7369 SMITH CLERK 7902 17-DEC-80 800 20
7499 ALLEN SALESMAN 7698 20-FEB-81 1600 300 30
7521 WARD SALESMAN 7698 22-FEB-81 1250 500 30
7566 JONES MANAGER 7839 02-APR-81 2975 20
7654 MARTIN SALESMAN 7698 28-SEP-81 1250 1400 30
7698 BLAKE MANAGER 7839 01-MAY-81 2850 30
7782 CLARK MANAGER 7839 09-JUN-81 2450 10
7788 SCOTT ANALYST 7566 19-APR-87 3000 20
7839 KING PRESIDENT 17-NOV-81 5000 10
7844 TURNER SALESMAN 7698 08-SEP-81 1500 0 30
7876 ADAMS CLERK 7788 23-MAY-87 1100 20
7900 JAMES CLERK 7698 03-DEC-81 950 30
7902 FORD ANALYST 7566 03-DEC-81 3000 20
7934 MILLER CLERK 7782 23-JAN-82 1300 10
14 rows selected.
5、将数据还原
可以插入回去,如闪回查询那样:insert into test10 select * from scott.test10 as of scn 1952615;
另外一种,可以用闪回表的方式:flashback table scott.test10 to scn 1952615;
注意,闪回表时要开启 row movement,不然要报错。
sys@TEST0910> flashback table scott.test10 to scn 1952615;
flashback table scott.test10 to scn 1952615
*
ERROR at line 1:
ORA-08189: cannot flashback the table because row movement is not enabled
sys@TEST0910> alter table scott.test10 enable row movement;
Table altered.
sys@TEST0910> flashback table scott.test10 to scn 1952615;
Flashback complete.
sys@TEST0910> select count(*) from scott.test10;
COUNT(*)
----------
14
1 row selected.
flashback database实验(使用lgmnr找出还原时间点)
过程:激活日志--删数据--查日志--logmnr--关库--不用转储--mount--flashback数据库(时间减一秒)--read only
开库看正确与否--关库--mount--开库resetlogs
DDL语句的恢复:类似不完全恢复,可以恢复到某个时间点,依赖flashback log.
logmnr依赖的是redo和归档。
flashback database一般用于ddl操作,
dml用flashback query或flashback table。因为他们读undo块
ddl语句不生成undo块
闪回数据库有限制:比如解决不了media recovery等(硬盘坏了等)。还是备份最重要。
1、使用logmnr来记录时间点,首先必须设置有一下参数
show parameter utl; logmnr需要的路径
sys@TEST0910> show parameter utl
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
create_stored_outlines string
utl_file_dir string /u01/app/oracle/utl
使用flashback database首先设置两个参数:
db_recovery_file_dest string /u01/app/oracle/fast_recovery_area
db_recovery_file_dest_size big integer 4122M
sys@TEST0910> show parameter recovery
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
db_recovery_file_dest string /u01/app/oracle/fast_recovery_
area
db_recovery_file_dest_size big integer 4122M
而且需要alter database flashback on
ys@TEST0910> select name,flashback_on from v$database;
NAME FLASHBACK_ON
-------------------------------------------------- ------------------
TEST0910 YES
1 row selected.
2、查询原数据
sys@TEST0910> select count(*) from scott.test12;
COUNT(*)
----------
14
1 row selected.
3.激活logmnr,注意,必须有此步骤,不然logmnr出不来数据
sys@TEST0910> ALTER DATABASE ADD SUPPLEMENTAL LOG DATA;
Database altered.
4、误操作DDL操作
sys@TEST0910> truncate table scott.test12;
Table truncated.
sys@TEST0910> select count(*) from scott.test12;
COUNT(*)
----------
0
1 row selected.
5、查询当前日志组和归档日志
sys@TEST0910> select * from v$log;
GROUP# THREAD# SEQUENCE# BYTES BLOCKSIZE MEMBERS ARC STATUS FIRST_CHANGE# FIRST_TIME NEXT_CHANGE# NEXT_TIME
---------- ---------- ---------- ---------- ---------- ---------- --- ---------------- ------------- ------------------- ------------ -------------------
1 1 7 52428800 512 1 YES INACTIVE 1977009 2013-09-23 15:04:56 1981677 2013-09-23 15:05:11
2 1 8 52428800 512 1 YES ACTIVE 1981677 2013-09-23 15:05:11 1985831 2013-09-23 15:05:41
3 1 9 52428800 512 1 NO CURRENT 1985831 2013-09-23 15:05:41 2.8147E+14
3 rows selected.
sys@TEST0910> select * from v$logfile;
GROUP# STATUS TYPE MEMBER IS_
---------- ------- ------- -------------------------------------------------- ---
2 ONLINE /u01/app/oracle/oradata/test0910/redo02.log NO
1 ONLINE /u01/app/oracle/oradata/test0910/redo01.log NO
3 ONLINE /u01/app/oracle/oradata/test0910/redo03.log NO
3 rows selected.
sys@TEST0910> select name from v$archived_log;
NAME
----------------------------------------------------------------------------------------------------
/u01/app/oracle/fast_recovery_area/TEST0910/archivelog/2013_09_23/o1_mf_1_6_94146rg7_.arc
/u01/app/oracle/fast_recovery_area/TEST0910/archivelog/2013_09_23/o1_mf_1_7_941477xf_.arc
/u01/app/oracle/fast_recovery_area/TEST0910/archivelog/2013_09_23/o1_mf_1_8_941485th_.arc
61 rows selected.
6、执行logmnr
EXECUTE DBMS_LOGMNR_D.BUILD(‘dictionary.ora‘, -
‘/u01/app/oracle/utl‘, -
DBMS_LOGMNR_D.STORE_IN_FLAT_FILE);
EXECUTE DBMS_LOGMNR.ADD_LOGFILE( -
LOGFILENAME => ‘/u01/app/oracle/oradata/test0910/redo03.log‘, -
OPTIONS => DBMS_LOGMNR.NEW);
EXECUTE DBMS_LOGMNR.ADD_LOGFILE( -
LOGFILENAME => ‘/u01/app/oracle/fast_recovery_area/TEST0910/archivelog/2013_09_23/o1_mf_1_6_94146rg7_.arc‘, -
OPTIONS => DBMS_LOGMNR.ADDFILE);
EXECUTE DBMS_LOGMNR.ADD_LOGFILE( -
LOGFILENAME => ‘/u01/app/oracle/fast_recovery_area/TEST0910/archivelog/2013_09_23/o1_mf_1_7_941477xf_.arc‘, -
OPTIONS => DBMS_LOGMNR.ADDFILE);
EXECUTE DBMS_LOGMNR.ADD_LOGFILE( -
LOGFILENAME => ‘/u01/app/oracle/fast_recovery_area/TEST0910/archivelog/2013_09_23/o1_mf_1_8_941485th_.arc‘, -
OPTIONS => DBMS_LOGMNR.ADDFILE);
EXECUTE DBMS_LOGMNR.START_LOGMNR( -
DICTFILENAME =>‘/u01/app/oracle/utl/dictionary.ora‘,OPTIONS => -
DBMS_LOGMNR.DDL_DICT_TRACKING);
EXECUTE DBMS_LOGMNR.END_LOGMNR;
logminer官方参考:http://docs.oracle.com/cd/E11882_01/server.112/e22490/logminer.htm#i1009063
以下为执行步骤
sys@TEST0910> EXECUTE DBMS_LOGMNR_D.BUILD(‘dictionary.ora‘, -
> ‘/u01/app/oracle/utl‘, -
> DBMS_LOGMNR_D.STORE_IN_FLAT_FILE);
LogMnr Dictionary Procedure started
LogMnr Dictionary File Opened
Procedure executed successfully - LogMnr Dictionary Created
PL/SQL procedure successfully completed.
sys@TEST0910> EXECUTE DBMS_LOGMNR.ADD_LOGFILE( -
> LOGFILENAME => ‘/u01/app/oracle/oradata/test0910/redo03.log‘, -
> OPTIONS => DBMS_LOGMNR.NEW);
PL/SQL procedure successfully completed.
sys@TEST0910> EXECUTE DBMS_LOGMNR.ADD_LOGFILE( -
> LOGFILENAME => ‘/u01/app/oracle/fast_recovery_area/TEST0910/archivelog/2013_09_23/o1_mf_1_8_941485th_.arc‘, -
> OPTIONS => DBMS_LOGMNR.ADDFILE);
PL/SQL procedure successfully completed.
sys@TEST0910> EXECUTE DBMS_LOGMNR.ADD_LOGFILE( -
> LOGFILENAME => ‘/u01/app/oracle/fast_recovery_area/TEST0910/archivelog/2013_09_23/o1_mf_1_7_941477xf_.arc‘, -
> OPTIONS => DBMS_LOGMNR.ADDFILE);
PL/SQL procedure successfully completed.
sys@TEST0910> EXECUTE DBMS_LOGMNR.ADD_LOGFILE( -
> LOGFILENAME => ‘/u01/app/oracle/fast_recovery_area/TEST0910/archivelog/2013_09_23/o1_mf_1_6_94146rg7_.arc‘, -
> OPTIONS => DBMS_LOGMNR.ADDFILE);
PL/SQL procedure successfully completed.
sys@TEST0910> EXECUTE DBMS_LOGMNR.START_LOGMNR( -
> DICTFILENAME =>‘/u01/app/oracle/utl/dictionary.ora‘,OPTIONS => -
> DBMS_LOGMNR.DDL_DICT_TRACKING);
PL/SQL procedure successfully completed.
7、设置session时间格式
sys@TEST0910> alter session set nls_date_format=‘yyyy-mm-dd hh24:mi:ss‘;
Session altered.
8、查询logmnr分析出的数据
sys@TEST0910> select username,scn,timestamp,sql_redo from v$logmnr_contents
2 where seg_name=‘TEST12‘;
USERNAME SCN TIMESTAMP SQL_REDO
------------------------------ ---------- ------------------- --------------------------------------------------
SYS 1989009 2013-09-23 15:09:53 truncate table scott.test12;
sys@TEST0910> EXECUTE DBMS_LOGMNR.END_LOGMNR;
PL/SQL procedure successfully completed.
8、mount状态下闪回
sys@TEST0910> shutdown immediate;
Database closed.
Database dismounted.
ORACLE instance shut down.
sys@TEST0910> startup mount;
ORACLE instance started.
Total System Global Area 2505338880 bytes
Fixed Size 2230952 bytes
Variable Size 587203928 bytes
Database Buffers 1895825408 bytes
Redo Buffers 20078592 bytes
Database mounted.
sys@TEST0910> flashback database to timestamp to_timestamp(‘2013-09-23 15:09:52‘,‘yyyy-mm-dd hh24:mi:ss‘);
此处注意,时间在logmnr查询出的基础上减1秒
Flashback complete.
9、让数据库置于read only,检查数据是否正确恢复
sys@TEST0910> alter database open read only;
Database altered.
sys@TEST0910> select count(*) from scott.test12;
COUNT(*)
----------
14
1 row selected.
10、重新启动数据库,使用resetlogs打开数据库
sys@TEST0910> shutdown immediate;
Database closed.
Database dismounted.
ORACLE instance shut down.
sys@TEST0910> startup mount;
ORACLE instance started.
Total System Global Area 2505338880 bytes
Fixed Size 2230952 bytes
Variable Size 587203928 bytes
Database Buffers 1895825408 bytes
Redo Buffers 20078592 bytes
Database mounted.
sys@TEST0910> alter database open
2 ;
alter database open
*
ERROR at line 1:
ORA-01589: must use RESETLOGS or NORESETLOGS option for database open
sys@TEST0910> alter database open resetlogs;
Database altered.
声明:以上内容来自用户投稿及互联网公开渠道收集整理发布,本网站不拥有所有权,未作人工编辑处理,也不承担相关法律责任,若内容有误或涉及侵权可进行投诉: 投诉/举报 工作人员会在5个工作日内联系你,一经查实,本站将立刻删除涉嫌侵权内容。