用户管理恢复 一、Oracle恢复的体系结构
二、恢复类型：实例恢复、介质恢复
1、实例恢复的前提是数据文件、控制文件和在线日志均没有损坏，完全自动
2、介质恢复：在文件有物理损坏的情况下，将其还原、恢复、最后使数据库得以正常打开的操作称为介质恢复，人工干预
三、恢复方式：完全恢复、不完全恢复
四、恢复技术：用户管理恢复和RMAN恢复
五、完全恢复案例
v$recover_file //哪个文件需要恢复 select file#,change# from v$recover_file;
v$log //在线日志 select min(first_change#) from v$log where status!='INACTIVE';
v$archived_log //归档日志 select min(first_change#) from v$archived_log;
v$recovery_log //需要哪些archvielog做恢复 1、*场景1:数据库关闭时的恢复(system,sysaux,undo,whole database):system表空间为例(冷备)
select * from v$dbfile;
shutdown immediate;
cp -rf /u01/app/oracle/oradata/ocp/* /backup/cold --冷备整个数据文件
startup
conn gyj/gyj
insert into t1 values(4,'gyj4');
commit;
alter system switch logfile; --切换日志
rm -rf system01.dbf --册除system系统表空间
shutdown abort; --异常关机
cp /backup/cold/system01.dbf /u01/app/oracle/oradata/ocp/ --用备份先做还原
startup; --会报错:ORA-01157，ORA-01110
recover datafile 1; --恢复
一直回车
alter database open; 2、*场景2:数据库打开时的恢复以用户数据的表空间为例(热备)
create tablespace tp5 datafile '/u01/app/oracle/oradata/ocp/tp5.dbf' size 10M;
conn gyj/gyj
create table t5 (id int,name varchar2(10)) tablespace tp5;
insert into t5 values(1,'gyj1');
commit;
alter tablespace tp5 begin backup; --热备
!cp /u01/app/oracle/oradata/ocp/tp5.dbf /backup/hot
alter tablespace tp5 end backup;
conn gyj/gyj
insert into t5 values(2,'gyj2');
commit;
rm -rf tp5.dbf;
alter database datafile 9 offline;
!cp /backup/hot/tp5.dbf /u01/app/oracle/oradata/ocp/ --用备份先做还原
recover datafile 9; --恢复
alter database datafile 11 online;
3、控制文件的恢复
select controlfile_sequence# from v$database;--当前控件文件记录的序列号
select hxfil as file#,FHCSQ from x$kcvfh;
select controlfile_change# from v$database; --控制自身的检查点
select file#,checkpoint_change# from v$datafile_header;
select checkpoint_change# from v$database; --数据库的检查点
select file#,checkpoint_change# from v$datafile; --从控制文件中读的scn
(1)***场景1:控制文件没有全坏,有一个是好的
rm -rf control02.ctl
shutdown abort;
cp control01.ctl control02.ctl
startup
(2)***场景2:所有控制文件都坏了,但损坏前有备份(二进制)
alter database backup controlfile to '/backup/control/control.bin';
alter system checkpoint;
rm -rf control*
shutdown abort;
cp /backup/control.bin control01.ctl
cp /backup/control.bin control02.ctl
cp /backup/control.bin control03.ctl
startup
select checkpoint_change# from v$database;
select checkpoint_change# from v$datafile;
select checkpoint_change# from v$datafile_header;
recover database using backup controlfile;
redologfile 一个个试
alter database open resetlogs; (3)*场景3:所有控制文件都坏了,但损坏前有备份(文本)
alter database backup controlfile to trace as '/backup/control/control.txt';
rm -rf control*
shutdown abort;
vi control.bak --重构一个新的控制文件
conn /as sysdba
@/backup/control/control.txt
recover database;
alter database open;
(4) *场景4:备份控制文件的恢复,控制文件中不包含数据文件,但联机日志中有
alter database backup controlfile to '/backup/control/control.bin';
create tablespace tp6 datafile '/u01/app/oracle/oradata/ocp/tp6.dbf' size 10M;
conn gyj/gyj
create table t6(id int,name varchar2(10)) tablespace tp6;
insert into t6 values(1,'gyj1');
commit;
rm -rf control*
shutdown abort;
cp /backup/control/control.bin control01.ctl
cp /backup/control/control.bin control02.ctl
cp /backup/control/control.bin control03.ctl
startup mount;
select file#,checkpoint_change# from v$datafile;
select file#,checkpoint_change# from v$datafile_header;
alter database open; //提示控制文件旧了
recover database using backup controlfile;
输入:/oradata/orcl/redo01.log --状态是current
提示有一个文件不识别
select file#,name from v$datafile;
alter database rename file '/u01/app/oracle/product/11g/dbs/UNNAMED00005' to '/u01/app/oracle/oradata/ocp/tp6.dbf'; --看上面提示或 看alert.log日志
recover database using backup controlfile;
输入:/oradata/orcl/redo01.log --一select * from v$log;
alter database open resetlogs;
二、不完全恢复案例
1、不完全恢复的动因
(1)被动恢复：原计划执行完全恢复，但是恢复的过程中发现recover命令需要的至少一个归档日志或未归档的ACTIVE/CURRENT在线日志损坏
(2)主动恢复：由于人为错误码或恶意篡改破坏了数据库内大量的信息（update,truncate） 2、不完全恢复：recover命令的四种情况
(1)精确到重做日志的基于日志序列号的“recover database until sequence”--仅适用RMAN
(2)精确到到重做日志的基于手动交互式的“recover database until cancel”--仅适用SQL*PLUS
(3)精确到重做记录的基于时间的 “recover database until time”--适用RMAN和SQL*PLUS
(4)精确到重做记录的基于SCN的“recover database until scn”--仅适用RMAN 3、基本时间的不完全恢复
(1)*场景１:误操作删除表，控制文件没损坏
shutdown immediate;
!cp /u01/app/oracle/oradata/ocp/* /backup/cold/ --做一个冷备
startup
conn gyj/gyj
insert into t1 values(3,'gyj3');
commit;
select to_char(sysdate,'yyyy-mm-dd:hh24:mi:ss') from dual;--恢复此时间之前的数据
insert into t1 values(4,'gyj4');
commit;
drop table t1; --模拟误操作
shutdown immediate;
rm -rf /u01/app/oracle/oradata/ocp/*.dbf --删除所有数据(不删控制文件和重做日志文件)
cp -rf /backup/cold/*.dbf /u01/app/oracle/oradata/ocp/ --restore所有数据
startup mount;
recover database until time '2013-03-29 09:49:05'; ---基于时间的不完全恢复
alter database open resetlogs; (2) *场景1:误操作删除表，控制文件也被损坏了
shutdown immediate;
!cp /u01/app/oracle/oradata/ocp/* /backup/cold/ --做一个冷备
startup
conn gyj/gyj
insert into t1 values(6,'gyj6');
commit;
select to_char(sysdate,'yyyy-mm-dd:hh24:mi:ss') from dual;--误操作前的时间
2012-06-22:17:09:08
insert into t1 values(7,'gyj7');
commit;
drop table t1; --模拟误操作
shutdown abort;
rm -rf /u01/app/oracle/oradata/ocp/*.dbf --删除数据文件
rm -rf /u01/app/oracle/oradata/ocp/*.ctl --删除控制文件,(不删除重做日志文件)
cp -rf /backup/cold/*.dbf /u01/app/oracle/oradata/ocp/ --还原所有数据
cp -rf /backup/cold/*.ctl /u01/app/oracle/oradata/ocp/ --还原控制文件
startup mount;
recover database until time '2012-10-31:11:47:15' using backup controlfile; --用了备份的控制文件
alter database open resetlogs; 4、基本Cancel的不完全恢复
(1)***场景1:写到数据文件,归档没保留
shutdown immediate;
!cp -rf /u01/app/oracle/oradata/ocp/* /backup/cold/ --重生做一个冷备 startup
conn gyj/gyj
insert into t1 values(4,'gyj4');
commit;
alter system switch logfile;
insert into t1 values(5,'gyj5');
commit;
select group#,status from v$log;
alter system checkpoint; --写到数据文件,归档没保留 rm -rf /u01/app/oracle/oradata/ocp/ --把所有的/oradata/orcl/ 全部删掉
shutdown abort;
cp /backup/cold/*.dbf /u01/app/oracle/oradata/ocp/ --controlfile,redolog都没了
cp /backup/cold/*.ctl /u01/app/oracle/oradata/ocp/
startup mount; ; recover database using backup controlfile until cancel; --用了备份的控制文件
--如归档存在,直接回车
--如归档不存在,输入cancel,因为redolog全删了
alter database open resetlogs;
alter tablespace temp add tempfile '/u01/app/oracle/oradata/bxdb/temp.dbf' size 10M;
select * from t; --数据少了一条
ALTER TABLESPACE TEMP ADD TEMPFILE '/u01/app/oracle/oradata/orcl/temp01.dbf' REUSE;
RMAN恢复 1、RMAN完全恢复
(1)*场景1:参数文件损坏
backup spfile;
--create pfile='/tmp/pfile.ora' from spfile;
shutdown immediate;
rm -rf $ORACLE_HOME/dbs/spfileocp.ora
startup pfile='/tmp/pfile.ora' nomount;
set dbid 43434343
restore spfile from autobackup;
restore spfile to '/temp/spfile.ora' from autobackup;
(2)*场景2:控制文件损坏
backup current controlfile;
rm -rf control*
startup nomount;
set dbid=2690543208;
restore controlfile from autobackup;--db_recovery_file_dest或$ORACLE_HOME/dbs(set dbid)
alter database mount;
recover database;
alter database open resetlogs; (3) *场景3 system表空间数据文件损坏
backup database;
rm -rf system01.dbf;
shutdown abort;
startup mount;
restore datafile 1;
recover datafile 1;
alter database open; (4)*场景4 undo表空间损坏
startup mount;
restore tablespace UNDOTBS1;
recover database;
alter database open; (5)*场景5:磁盘损坏
rm -rf tp1.dbf
select name,file# from v$datafile;
run{ shutdown abort;
startup mount;
set newname for datafile 6 to '/u01/app/oracle/oradata/ocm/tp1.dbf';
restore datafile 6;
switch datafile 6 ;
recover database;
alter database open;} (6)*场景6:普通数据文件损坏
rm -rf tp1.dbf 方法一
run{
shutdown abort;
startup mount;
restore datafile 6;
recover datafile 6;
alter database open;} 方法二
sql 'alter database datafile 6 offline';
restore datafile 6;
recover datafile 6;
sql 'alter database datafile 6 online';} 方法三
shutdown abort;
startup mount;
sql 'alter database datafile 6 offline';
alter database open;
restore datafile 6;
recover datafile 6;
sql 'alter database datafile 6 online';} (7)控制文件无损的情况下，只读文件损坏
*场景1:只读数据文件丢失
run{
shutdown abort;
startup mount;
restore datafile 5;
alter database open;} *场景2:只读数据文件在实例运行时丢失
run{
sql 'alter database datafile 5 offline';
restore datafile 5;
sql 'alter database datafile 5 online';} *场景3:只读数据文件内部数据块损坏
run{
sql 'alter database datafile 5 offline';
restore datafile 5 force;
sql 'alter database datafile 5 online';} *场景4:起先数据文件备份时是只读状态，后来其表空间改为读写状态，但没有备份，现数据文件坏了
run{
sql 'alter database datafile 6 offline';
restore datafile 6;
recover datafile 6;
sql 'alter database datafile 6 online';} (8)控制文件损坏的情况下，只读文件损坏
分析只读表空间在控制文件的情况
alter database backup controlfile to trace as '/backup/control.bak01';
alter tablespace read only;
alter database backup controlfile to trace as '/backup/control.bak02';
diff control.bak01 control.bak02 --比较两个备份的控制文件 alter database rename file '' to ''; --只读表空间在重构时不考虑,要在数据库打开后,重命名数据文件
alter database tp1 online; *场景1:只读数据文件连同控件文件一并损坏或丢失
run{
shutdown abort;
startup nomount;
restore controlfile from autobackup;
alter database mount;
restore datafile 6;
recover database;
alter database open resetlogs;}
*场景2:在数据文件为只读状态时，对该文件和控制文件进行了备份，随且对该表空间改为读写状态，但该文件和控制文件均未再有过备份，此时该数据文件连同控件文件一并损坏或丢失
run{
shutdown abort;
startup nomount;
restore controlfile from '/backup/rman/full_0cnnlmfd_1_1';
mount database;
restore datafile 6;
recover database;
recover database; --注意再次执行
alter database open resetlogs;}
(9)表空间的恢复
[过程：全备->建立表空间->在此表空间上进行业务->此表空间被 DROP->恢复此表空间
[oracle@redhat55 ~]$ rman target/
Recovery Manager:Release 10.2.0.1.0 - Production on 星期二 4月 30 23:23:19 2013
Copyright (c) 1982,2005, Oracle. All rights reserved.
connected to targetdatabase: ORCL (DBID=1341748519)
RMAN> list backup;
using targetdatabase control file instead of recovery catalog
List of Backup Sets
BS Key Type LV Size Device Type Elapsed Time CompletionTime
1 Full 30.83M DISK 00:01:00 2013-04-30 23:21:46
BP Key: 1 Status: AVAILABLE Compressed: YES Tag: TAG20130430T232045
Piece Name:/disk1/backup/orcl/02o8dnae_1_1.bak
List of Datafiles in backup set 1
File LV Type Ckp SCN Ckp Time Name
2 Full 573612 2013-04-3023:20:46 /u01/app/oracle/oradata/orcl/undotbs01.dbf
3 Full 573612 2013-04-3023:20:46 /u01/app/oracle/oradata/orcl/sysaux01.dbf
5 Full 573612 2013-04-3023:20:46 /u01/app/oracle/oradata/orcl/example01.dbf
2 Full 80.84M DISK 00:01:19 2013-04-30 23:22:05
BP Key: 2 Status: AVAILABLE Compressed: YES Tag: TAG20130430T232045
Piece Name:/disk1/backup/orcl/01o8dnae_1_1.bak
List of Datafiles in backup set 2
1 Full 573611 2013-04-3023:20:46 /u01/app/oracle/oradata/orcl/system01.dbf
4 Full 573611 2013-04-3023:20:46 /u01/app/oracle/oradata/orcl/users01.dbf
3 Full 6.80M DISK 00:00:00 2013-04-30 23:22:07
BP Key: 3 Status: AVAILABLE Compressed: NO Tag: TAG20130430T232207
Piece Name: /disk1/backup/orcl/c-1341748519-20130430-00.clt
Control File Included: Ckp SCN: 573653 Ckp time: 2013-04-30 23:22:07
SPFILE Included: Modification time:2013-04-30 23:17:26
RMAN>
[oracle@redhat55 ~]$ sqlplus /as sysdba
SQL*Plus: Release10.2.0.1.0 - Production on 星期二 4月 30 23:24:35 2013
Connected to:
Oracle Database 10gEnterprise Edition Release 10.2.0.1.0 - 64bit Production
With thePartitioning, OLAP and Data Mining options
SQL>
SQL> create tablespace tbsdatafile '/u01/app/oracle/oradata/orcl/tbs.dbf' size 5m;
Tablespace created.
SQL> create table t (id int)tablespace tbs;
Table created.
SQL> insert into t values(1);
1 row created.
SQL> commit;
Commit complete.
SQL> select * from t;
ID
1
SQL> select file_name fromdba_data_files;
FILE_NAME
/u01/app/oracle/oradata/orcl/users01.dbf
/u01/app/oracle/oradata/orcl/sysaux01.dbf
/u01/app/oracle/oradata/orcl/undotbs01.dbf
/u01/app/oracle/oradata/orcl/system01.dbf
/u01/app/oracle/oradata/orcl/example01.dbf
/u01/app/oracle/oradata/orcl/tbs.dbf
6 rows selected.
SQL> select current_scn fromv$database;
CURRENT_SCN
573950 // 恢复时恢复到这个SCN号
SQL> alter system archive logcurrent;
System altered.
SQL> drop tablespace tbsINCLUDING CONTENTS and datafiles;
Tablespace dropped.
SQL> select file_namefrom dba_data_files; // 表空间对应的数据文件已经没有了
SQL> shutdown abort;
ORACLE instance shutdown.
SQL> startup nomount;
ORACLE instancestarted.
Total System GlobalArea 1224736768 bytes
Fixed Size 2020384 bytes
Variable Size 318770144 bytes
DatabaseBuffers 889192448 bytes
Redo Buffers 14753792 bytes
SQL> exit
Disconnected fromOracle Database 10g Enterprise Edition Release 10.2.0.1.0 - 64bit Production
Recovery Manager:Release 10.2.0.1.0 - Production on 星期二 4月 30 23:29:31 2013
connected to targetdatabase: orcl (not mounted)
RMAN> restore controlfilefrom '/disk1/backup/orcl/c-1341748519-20130430-00.clt';
Starting restore at2013-04-30 23:29:54
allocated channel:ORA_DISK_1
channel ORA_DISK_1:sid=156 devtype=DISK
channel ORA_DISK_1:restoring control file
channel ORA_DISK_1:restore complete, elapsed time: 00:00:02
outputfilename=/u01/app/oracle/oradata/orcl/control01.ctl
outputfilename=/u01/app/oracle/oradata/orcl/control02.ctl
outputfilename=/u01/app/oracle/oradata/orcl/control03.ctl
Finished restore at2013-04-30 23:29:57
RMAN> shutdown abort;
Oracle instance shutdown
RMAN> startup mount;
connected to targetdatabase (not started)
Oracle instancestarted
database mounted
RMAN> restore database;
Starting restore at2013-04-30 23:30:30
Starting implicitcrosscheck backup at 2013-04-30 23:30:30
channel ORA_DISK_1:sid=157 devtype=DISK
allocated channel:ORA_DISK_2
channel ORA_DISK_2:sid=155 devtype=DISK
Crosschecked 2objects
Finished implicitcrosscheck backup at 2013-04-30 23:30:33
Starting implicitcrosscheck copy at 2013-04-30 23:30:33
using channelORA_DISK_1
using channelORA_DISK_2
Finished implicitcrosscheck copy at 2013-04-30 23:30:33
searching for allfiles in the recovery area
cataloging files...
no files cataloged
channel ORA_DISK_1:starting datafile backupset restore
channel ORA_DISK_1:specifying datafile(s) to restore from backup set
restoring datafile00002 to /u01/app/oracle/oradata/orcl/undotbs01.dbf
restoring datafile00003 to /u01/app/oracle/oradata/orcl/sysaux01.dbf
restoring datafile00005 to /u01/app/oracle/oradata/orcl/example01.dbf
channel ORA_DISK_1:reading from backup piece /disk1/backup/orcl/02o8dnae_1_1.bak
channel ORA_DISK_2:starting datafile backupset restore
channel ORA_DISK_2:specifying datafile(s) to restore from backup set
restoring datafile00001 to /u01/app/oracle/oradata/orcl/system01.dbf
restoring datafile00004 to /u01/app/oracle/oradata/orcl/users01.dbf
channel ORA_DISK_2:reading from backup piece /disk1/backup/orcl/01o8dnae_1_1.bak
channel ORA_DISK_1:restored backup piece 1
piecehandle=/disk1/backup/orcl/02o8dnae_1_1.bak tag=TAG20130430T232045
channel ORA_DISK_1:restore complete, elapsed time: 00:00:46
channel ORA_DISK_2:restored backup piece 1
piecehandle=/disk1/backup/orcl/01o8dnae_1_1.bak tag=TAG20130430T232045
channel ORA_DISK_2:restore complete, elapsed time: 00:01:11
Finished restore at2013-04-30 23:31:45
RMAN> run{
2> set until scn 573950;
3> recover database;
4> }
executing command:SET until clause
Starting recover at2013-04-30 23:33:37
starting mediarecovery
archive logfilename=/disk1/backup/orcl/1_3_813967785.arc thread=1 sequence=3
creating datafilefno=6 name=/u01/app/oracle/oradata/orcl/tbs.dbf
media recoverycomplete, elapsed time: 00:00:03
Finished recover at2013-04-30 23:33:43
RMAN> sql'alter database openresetlogs';
sql statement: alterdatabase open resetlogs
RMAN> exit
Recovery Managercomplete.
SQL*Plus: Release10.2.0.1.0 - Production on 星期二 4月 30 23:34:03 2013
/u01/app/oracle/oradata/orcl/tbs.dbf // 表空间已经恢复
总结：表空间在没有备份的情况下被DROP，此时若想恢复被DROP的表空间，必须使用DROP之前的备份及控制文件，然后利用归档日志再将数据库恢复到DROP之间的点，表空间会被归档日志里重新创建。
二、RMAN不完全恢复
(1)*场景1:基本时间的不完全恢复
startup
conn gyj/gyj
insert into t1 values(6,'gyj6');
commit;
select to_char(sysdate,'yyyy-mm-dd:hh24:mi:ss') from dual;
insert into t1 values(9,'gyj9');
commit;
drop table t1; --误操作
run{
shutdown abort;
startup mount;
set until time "to_date('2013-04-14:11:42:18','yyyy-mm-dd:hh24:mi:ss')";
restore database;
recover database;
alter database open resetlogs;}
(2)*场景2:基于sequence的不完全恢复
conn gyj/gyj
insert into t1 values(9,'gyj9');
commit;
alter system switch logfile;
alter system switch logfile;
alter system switch logfile;
insert into t1 values(10,'gyj10');
commit;
alter system switch logfile;
alter system switch logfile;
insert into t1 values(11,'gyj11');
alter system switch logfile;
rman target /
run{
shutdown abort;
startup mount;
set until sequence 10 thread 1;
restore database;
recover database;
alter database open resetlogs;}
总结 一、关于recover database命令
（1）SQL>recover database;
该命令用来对所有数据文件进行恢复，并且只能使用保存在文件系统上的归档日志及在线日志。使用此命令的前提是控制文件不可以是还原或重建得来的。
Oracle会以当前controlfile所记录的SCN为准，利用archive log和 redo log的redo entry, 把相关的datafile 的 block恢复到“当前controlfile所记录的SCN” （2）SQL>recover database using backup controlfile;
该命令用来对所有数据文件及控制文件进行恢复，并且只使用保存在文件系统上的归档及在线日志。
Oracle需要把数据恢复到比当前controlfile所记录的SCN还要靠后的位置（比如说，control file是backup controlfile , 或者 controlfile是根据trace create的。）,这时候，就需要用using backup controlfile. 恢复就不会受“当前controlfile所记录的SCN”的限制。 （3）RMAN>recover database;
该命令用来对所有数据文件及控制文件进行恢复，并且可以使用增量备份、备份中的和文件系统上的归档日志,以及在线日志。 二、关于resetlogs命令
（1）用备份的控制文件做恢复必须要resetlogs，使用resetlogs的原因是recover命令只能修复控制文件中数据库物理结构信息，而无法修改控制文件中的当前重做日志的序列号等信息
（2）不完全恢复的resetlogs
三、数据库的化身－－Incarnation
v$database_incarnation
RMAN>list incarnation
![【MySQL数据库】数据库索引事务1.索引2.事务[图]](data:image/gif;base64,R0lGODlhAQABAIAAAP///wAAACwAAAAAAQABAAACAkQBADs=)