数据库恢复技术详解：基于时间、SCN和日志序列号的不完全恢复
只适用于archivelog模式，只能在mount状态下完成。
基于时间恢复
基于时间恢复是指当出现用户错误（例如误删除表、误截断表）时，恢复到指定时间点的恢复。
示例：
模拟误截断表t_user。
在终端设置环境变量nls_date_format指定日期时间格式。
执行rman,启动数据库到mount状态。
使用set until time命令指定要恢复到的时间点。
转储、恢复数据库，并使用resetlogs选项打开数据库。
--模拟误截断表t_user。 SQL> host date 2011年 08月 29日 星期一 21:01:29 CST SQL> truncate table t_user; Table truncated. --在终端设置环境变量nls_date_format指定日期时间格式。 [oracle@localhost ~]$ export NLS_DATE_FORMAT='YYYY-MM-DD HH24:MI:SS' --恢复 [oracle@localhost ~]$ rman target sys/oracle@oralife nocatalog RMAN> run { 2> startup force mount; 3> set until time='2011-08-29 21:01:00'; 4> restore database; 5> recover database; 6> sql 'alter database open resetlogs'; 7> } --已恢复 SQL> conn sys/oracle@oralife as sysdba Connected. SQL> select count(*) from t_user; COUNT(*) ---------- 2 在实际环境下，应该使用LogMiner确定误操作时间点。在执行了不完全恢复之后，推荐删除早期所有备份，重新备份数据库。如：
run {
delete noprompt backup;
delete noprompt copy;
backup database format='/oracle/10g/oracle/rman/%d_%s.dbf';
sql 'alter system archive log current';
}
基于SCN恢复
基于SCN恢复是指当出现用户错误（例如误删除表、误截断表）时，恢复到指定SCN点的恢复。
模拟误删除表t_user。
执行rman,启动数据库到mount状态。
使用set until scn命令指定要恢复到的scn点。
转储、恢复数据库，并使用resetlogs选项打开数据库。 --模拟误删除表t_user。 SQL> select count(*) from t_user; COUNT(*) ---------- 2 SQL> insert into t_user select 'spring mvc_' from dual; 1 row created. SQL> commit; Commit complete. SQL> alter system switch logfile; System altered. SQL> alter system checkpoint; System altered. SQL> select current_scn from v$database; CURRENT_SCN ----------- 1214281 SQL> drop table t_user; Table dropped. --恢复 [oracle@localhost ~]$ rman target sys/oracle@oralife nocatalog RMAN> run { 2> startup force mount; 3> set until scn=1214281; 4> restore database; 5> recover database; 6> sql 'alter database open resetlogs'; 7> } --已恢复 SQL> select count(*) from t_user; COUNT(*) ---------- 3 在实际环境下，应该使用LogMiner确定误操作SCN点。
在执行了不完全恢复之后，推荐删除早期所有备份，重新备份数据库。
基于日志序列号恢复
基于日志序列号恢复是指恢复数据库到指定日志序列号的状态。
在执行了不完全恢复之后，推荐删除早期所有备份，重新备份数据库。
基于备份控制文件恢复
基于备份控制文件恢复是指使用备份控制文件恢复数据库的过程。当误删除了表空间或数据库所有控制文件全部损坏时，可以使用这种恢复方法。
模拟误删除了test表空间，因为当前控制文件没有包含该表空间的信息，所以必须使用备份控制文件恢复被误删除的表空间。如果没有使用恢复目录，必须激活控制文件自动备份，否则将不转储控制文件备份。
模拟误删除test表空间。
查看alert日志文件，确定操作时间。
使用rman启动数据库到mount状态。
设置数据库id。
转储控制文件，恢复数据库。
恢复失败，失败原因:restore controlfile from autobackup;用于转储自动备份的最新的控制文件，即删除表空间rtest之后自动备份的控制文件。
解决办法：使用restore controlfile from autobackup until time "to_date('2011-08-31 21:26:16','yyyy-mm-dd hh24:mi:ss')"; 查看alert日志文件确定删除表空间rtest之前自动备份控制文件时间，也就是删除表空间rtest之前的时间点。
SQL> create tablespace rtest DATAFILE '/oracle/10g/oracle/product/10.2.0/oradata/oralife/rtest.dbf' size 10M; Tablespace created. SQL> create table r_test (text varchar(20)) tablespace rtest; Table created. SQL> insert into r_test select '00001' from dual; 1 row created. SQL> commit; Commit complete. SQL> alter system switch logfile; System altered. SQL> alter system checkpoint; System altered. --删除早期所有备份，重新备份数据库 RMAN> run { delete noprompt backup; delete noprompt copy; backup database format='/oracle/10g/oracle/rman/%d_%s.dbf'; sql 'alter system archive log current'; } --删除表空间与对应的数据文件 SQL> insert into r_test select '00002' from dual; 1 row created. SQL> commit; Commit complete. SQL> alter system switch logfile; System altered. SQL> alter system checkpoint; System altered. SQL> select count(*) from r_test; COUNT(*) ---------- 2 SQL> drop tablespace rtest including contents and datafiles; Tablespace dropped. SQL> select dbid from v$database; DBID ---------- 3898974595 SQL> select count(*) from r_test; select count(*) from r_test * ERROR at line 1: ORA-00942: table or view does not exist --查看alert日志，以确定操作时间，Wed Aug 31 21:26:17 2011 SQL> host tail -n 20 /oracle/10g/oracle/product/10.2.0/db_1/admin/oralife/bdump/alert_oralife.log Current log# 4 seq# 4 mem# 0: /oracle/10g/oracle/product/10.2.0/oradata/oralife/redo04.log Current log# 4 seq# 4 mem# 1: /oracle/10g/oracle/product/10.2.0/oradata/oralife/redo04_add.log Wed Aug 31 21:25:26 2011 Thread 1 advanced to log sequence 5 Current log# 2 seq# 5 mem# 0: /oracle/10g/oracle/product/10.2.0/oradata/oralife/redo02.log Current log# 2 seq# 5 mem# 1: /oracle/10g/oracle/product/10.2.0/oradata/oralife/redo02_add.log Wed Aug 31 21:26:17 2011 drop tablespace rtest including contents and datafiles Wed Aug 31 21:26:18 2011 Deleted file /oracle/10g/oracle/product/10.2.0/oradata/oralife/rtest.dbf Starting control autobackup Wed Aug 31 21:26:18 2011 Errors in file /oracle/10g/oracle/product/10.2.0/db_1/admin/oralife/udump/oralife_ora_3646.trc: Wed Aug 31 21:26:18 2011 Errors in file /oracle/10g/oracle/product/10.2.0/db_1/admin/oralife/udump/oralife_ora_3646.trc: Wed Aug 31 21:26:18 2011 Errors in file /oracle/10g/oracle/product/10.2.0/db_1/admin/oralife/udump/oralife_ora_3646.trc: Control autobackup written to DISK device handle '/oracle/10g/oracle/product/10.2.0/db_1/flash_recovery_area/ORALIFE/autobackup/2011_08_31/o1_mf_s_760656378_75wfmtcm_.bkp' Completed: drop tablespace rtest including contents and datafiles --恢复控制文件 [oracle@localhost ~]$ export NLS_DATE_FORMAT='YYYY-MM-DD HH24:MI:SS' [oracle@localhost ~]$ rman target sys/oracle@oralife nocatalog RMAN> startup force nomount; RMAN> set dbid=3898974595; --与下面set until time（即删除表空间rtest的操作时间)相比少了1秒， 以转储正确的控制文件， --如果这里设置的时间 与alert日志中删除表空间的操作时间相同，会转储最新的控制文件（没有包含表空间rtest信息)导致恢复失败。 RMAN> restore controlfile from autobackup until time "to_date('2011-08-31 21:26:16','yyyy-mm-dd hh24:mi:ss')"; Starting restore at 2011-08-31 21:35:29 using channel ORA_DISK_1 recovery area destination: /oracle/10g/oracle/product/10.2.0/db_1/flash_recovery_area database name (or database unique name) used for search: ORALIFE channel ORA_DISK_1: autobackup found in the recovery area channel ORA_DISK_1: autobackup found: /oracle/10g/oracle/product/10.2.0/db_1/flash_recovery_area/ORALIFE/autobackup/2011_08_31/o1_mf_s_760656297_75wfk9gc_.bkp channel ORA_DISK_1: control file restore from autobackup complete output filename=/oracle/10g/oracle/product/10.2.0/oradata/oralife/control01.ctl output filename=/oracle/10g/oracle/product/10.2.0/oradata/oralife/control02.ctl output filename=/oracle/10g/oracle/product/10.2.0/oradata/oralife/control03.ctl Finished restore at 2011-08-31 21:35:33 RMAN> alter database mount; RMAN> run { set until time='2011-08-31 21:26:17'; restore database; recover database; sql 'alter database open resetlogs'; 6> } executing command: SET until clause Starting restore at 2011-08-31 21:36:43 Starting implicit crosscheck backup at 2011-08-31 21:36:43 allocated channel: ORA_DISK_1 channel ORA_DISK_1: sid=156 devtype=DISK Crosschecked 1 objects Finished implicit crosscheck backup at 2011-08-31 21:36:44 Starting implicit crosscheck copy at 2011-08-31 21:36:44 using channel ORA_DISK_1 Finished implicit crosscheck copy at 2011-08-31 21:36:44 searching for all files in the recovery area cataloging files... cataloging done List of Cataloged Files ======================= File Name: /oracle/10g/oracle/product/10.2.0/db_1/flash_recovery_area/ORALIFE/archivelog/2011_08_31/o1_mf_1_4_75wfl6w2_.arc File Name: /oracle/10g/oracle/product/10.2.0/db_1/flash_recovery_area/ORALIFE/archivelog/2011_08_31/o1_mf_1_3_75wfkc9w_.arc File Name: /oracle/10g/oracle/product/10.2.0/db_1/flash_recovery_area/ORALIFE/autobackup/2011_08_31/o1_mf_s_760656297_75wfk9gc_.bkp File Name: /oracle/10g/oracle/product/10.2.0/db_1/flash_recovery_area/ORALIFE/autobackup/2011_08_31/o1_mf_s_760656378_75wfmtcm_.bkp using channel ORA_DISK_1 channel ORA_DISK_1: starting datafile backupset restore channel ORA_DISK_1: specifying datafile(s) to restore from backup set restoring datafile 00001 to /oracle/10g/oracle/product/10.2.0/oradata/oralife/system01.dbf restoring datafile 00002 to /oracle/10g/oracle/product/10.2.0/oradatabak/undotbs01.dbf restoring datafile 00003 to /oracle/10g/oracle/product/10.2.0/oradatabak/sysaux01.dbf restoring datafile 00004 to /oracle/10g/oracle/product/10.2.0/oradata/oralife/user01.dbf restoring datafile 00005 to /oracle/10g/oracle/product/10.2.0/oradata/oralife/example01.dbf restoring datafile 00006 to /oracle/10g/oracle/product/10.2.0/oradata/oralife/rtest.dbf --转储了表空间rtest对应的数据文件 channel ORA_DISK_1: reading from backup piece /oracle/10g/oracle/rman/ORALIFE_98.dbf channel ORA_DISK_1: restored backup piece 1 piece handle=/oracle/10g/oracle/rman/ORALIFE_98.dbf tag=TAG20110831T212421 channel ORA_DISK_1: restore complete, elapsed time: 00:00:35 Finished restore at 2011-08-31 21:37:20 Starting recover at 2011-08-31 21:37:20 using channel ORA_DISK_1 starting media recovery archive log thread 1 sequence 3 is already on disk as file /oracle/10g/oracle/product/10.2.0/db_1/flash_recovery_area/ORALIFE/archivelog/2011_08_31/o1_mf_1_3_75wfkc9w_.arc archive log thread 1 sequence 4 is already on disk as file /oracle/10g/oracle/product/10.2.0/db_1/flash_recovery_area/ORALIFE/archivelog/2011_08_31/o1_mf_1_4_75wfl6w2_.arc archive log thread 1 sequence 5 is already on disk as file /oracle/10g/oracle/product/10.2.0/oradata/oralife/redo02.log archive log filename=/oracle/10g/oracle/product/10.2.0/db_1/flash_recovery_area/ORALIFE/archivelog/2011_08_31/o1_mf_1_3_75wfkc9w_.arc thread=1 sequence=3 archive log filename=/oracle/10g/oracle/product/10.2.0/db_1/flash_recovery_area/ORALIFE/archivelog/2011_08_31/o1_mf_1_4_75wfl6w2_.arc thread=1 sequence=4 archive log filename=/oracle/10g/oracle/product/10.2.0/oradata/oralife/redo02.log thread=1 sequence=5 media recovery complete, elapsed time: 00:00:03 Finished recover at 2011-08-31 21:37:25 sql statement: alter database open resetlogs --查看，已恢复 [oracle@localhost ~]$ sqlplus sys/oracle@oralife as sysdba SQL> select count(*) from r_test; COUNT(*) ---------- 2
扫一扫
543
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
