一、不同场景控制文件损坏时的恢复方法
场景 | 恢复方法 | 恢复条件 | |
其中一个控制文件损坏 | 1.1 拷贝冗余的控制文件 | 1、具有多路冗余控制文件镜像 | |
1.2 修改control_files参数去除损坏文件 | 同上、但不推荐该方法进行恢复 | ||
所有的控制文件损坏 | 有备份 | 2.1 通过rman备份控制文件进行完全恢复 | 1、通过rman备份了控制文件 |
2.2 通过rman备份控制文件进行不完全恢复 | 1、通过rman备份了控制文件 | ||
2.3 通过trace备份控制文件进行完全恢复 | 1、通过trace备份了控制文件 | ||
2.4 通过trace备份控制文件进行不完全恢复 | 1、通过trace备份了控制文件 | ||
无备份 | 2.5 通过手工重建控制文件进行恢复(noresetlogs) | 1、无有效的备份控制文件 | |
2.6 通过手工重建控制文件进行恢复(resetlogs) | 1、无有效的备份控制文件 | ||
2.7 通过SNAPSHOT CONTROLFILE文件进行恢复 | 此处为记录一个恢复控制文件的方法(不常使用) |
二、不同场景控制文件损坏的恢复思路及演示
2.1 其中一个控制文件损坏恢复思路
A. shutdown abort 关闭数据库(控制文件损坏后不能正常关闭数据库,只能使用abort强制关闭)
B. 拷贝其中一个完好的控制文件(推荐)或者修改control_files参数去除损坏文件(不推荐)
C. startup启动数据库
2.2 其中一个控制文件损坏恢复演示
SQL> show parameter control_files NAME TYPE VALUE ---------------------------------------------------------- ------------------------------ control_files string /u/app/oracle/oradata/racdg/control01.ctl,/u/app/oracle/ora data/racdg/control02.ctl SQL> !mv/u/app/oracle/oradata/racdg/control02.ctl /u/app/oracle/oradata/racdg/control02.ctl.bak SQL> shutdown immediate; --无正常关闭数据库,需要abort强制关闭 ORA-00210: cannot open the specifiedcontrol file ORA-00202: control file:'/u/app/oracle/oradata/racdg/control02.ctl' ORA-27041: unable to open file Linux-x86_64 Error: 2: No such file ordirectory Additional information: 3 SQL> shutdown abort; --只能shutdown abort强制关闭 ORACLE instance shut down. SQL> startup; --启动数据库 ORACLE instance started. ORA-00205: error in identifying controlfile, check alert log for more info --启动报错,找不到指定的控制文件,在告警日志文件alert_sid.log中可以看到如下报错: ORA-00210: cannot open the specifiedcontrol file ORA-00202: control file:'/u/app/oracle/oradata/racdg/control02.ctl' SQL> !pwd /u/app/oracle/oradata/racdg SQL> !cp control01.ctl control02.ctl --或者: SQL>alter system set control_files=’ /u/app/oracle/oradata/racdg/control01.ctl’scope=both; --不建议,因为这样一来就只剩一份控制文件了,起不到冗余作用,再此次发生丢失,恢复就变得麻烦了。 SQL> alter database mount; Database altered.
使用ASM存放控制文件,其中一个控制文件损坏恢复演示
SQL> show parameter control_files; NAME TYPE VALUE ----------------------------------------------- ------------------------------ control_files string +DATA/racdb/controlfile/current.256.935676497, +DATA/racdb/c ontrolfile/current.478.957977179 SQL> shutdown abort; --这里我是为了演示才直接关闭数据的,因为在线情况下asm的文件是不能删除的 ORACLE instance shut down. SQL> !su - grid -c "asmcmd -p rm +DATA/racdb/controlfile/current.478.957977179" Password: SQL> startup nomount; --将数据库启动到nomount状态 ORACLE instance started. SQL> !rman target / --这里使用rman来恢复,当然使用之前的两种方法拷贝和去除也是可以的 Recovery Manager: Release 11.2.0.4.0 -Production on Sat Oct 21 16:53:15 2017 Copyright (c) 1982, 2011, Oracle and/or itsaffiliates. All rights reserved. connected to target database: RACDB (notmounted) RMAN> restore controlfile from'+DATA/racdb/controlfile/current.256.935676497'; --其实这里就是拷贝的方式 Starting restore at 2017/10/21 16:54:18 using target database control file insteadof recovery catalog allocated channel: ORA_DISK_1 channel ORA_DISK_1: SID=143 instance=racdb1device type=DISK channel ORA_DISK_1: copied control filecopy output filename=+DATA/racdb/controlfile/current.256.935676497 output filename=+DATA/racdb/controlfile/current.478.957977659 Finished restore at 2017/10/21 16:54:22 RMAN> sql 'alter database mount'; sql statement: alter database mount released channel: ORA_DISK_1
说明一下,这里我分了文件系统和ASM存储存放控制文件的情况,其恢复原理都是一样的,无论是在线或者关闭数据库后损坏了其中部分控制文件,恢复方法都是一样的。
2.3所有的控制文件损坏,有备份场景
2.3.1通过rman备份控制文件进行完全恢复思路
A. 若在线损坏,shutdown abort关闭数据库;若关闭数据库后损坏,到第二步
B. startup nomount启动数据库到nomount状态。
C. 使用rman从备份中恢复控制文件:restorecontrolfile from '/path';
D. alter database mount;启动数据库到mount状态
E. recover database using backupcontrolfile until cancel; 应用归档文件
F. 再执行recover databaseusing backup controlfile; 应用未归档的redo文件
G. alter database open resetlogs; 以resetlogs的方式打开数据库
2.3.2通过rman备份控制文件进行完全恢复演示
SQL> show parameter control_files; NAME TYPE VALUE ---------------------------------------------------------- ------------------------------ control_files string /u/app/oracle/oradata/racdg/control01.ctl,/u/app/oracle/ora data/racdg/control02.ctl, /u/app/oracle/oradata/racdg/control03 .ctl SQL> !rm control01.ctl control02.ctlcontrol03.ctl --删除所有的控制文件,模拟控制文件丢失 SQL> shutdown abort; --强制关闭数据库 ORACLE instance shut down. SQL> startup; --启动数据到nomount状态 ORACLE instance started. ORA-00205: error in identifying controlfile, check alert log for more info --报错,同时查看告警日志,会发现所有控制文件都已经丢失。 SQL> !rman target / --此前有做过RMAN控制文件备份,使用rman进行恢复 Recovery Manager: Release 11.2.0.4.0 -Production on Sat Oct 21 17:24:11 2017 Copyright (c) 1982, 2011, Oracle and/or itsaffiliates. All rights reserved. connected to target database: RACDG (notmounted) RMAN> restore controlfile from'/u/app/oracle/zwdir/RACDG_20171021_0fshj32i_1_1.ctl'; --恢复控制文件 Starting restore at 2017/10/21 17:24:57 using target database control file insteadof recovery catalog allocated channel: ORA_DISK_1 channel ORA_DISK_1: SID=135 devicetype=DISK channel ORA_DISK_1: restoring control file channel ORA_DISK_1: restore complete,elapsed time: 00:00:03 output filename=/u/app/oracle/oradata/racdg/control01.ctl output filename=/u/app/oracle/oradata/racdg/control02.ctl output filename=/u/app/oracle/oradata/racdg/control03.ctl Finished restore at 2017/10/21 17:25:01 RMAN> sql 'alter database mount'; --将数据库启动到mount状态 sql statement: alter database mount released channel: ORA_DISK_1 RMAN> exit Recovery Manager complete. SQL> recover database using backupcontrolfile until cancel; ORA-00279: change 1041001 generated at10/21/2017 16:04:25 needed for thread 1 ORA-00289: suggestion :/u/app/oracle/oradata/arch/1_6_957480185.dbf ORA-00280: change 1041001 for thread 1 isin sequence #6 Specify log: {<RET>=suggested |filename | AUTO | CANCEL} auto ORA-00279: change 1041011 generated at10/21/2017 16:20:36 needed for thread 1 ORA-00289: suggestion :/u/app/oracle/oradata/arch/1_7_957480185.dbf ORA-00280: change 1041011 for thread 1 isin sequence #7 ORA-00278: log file'/u/app/oracle/oradata/arch/1_6_957480185.dbf' no longer needed for thisrecovery ORA-00279: change 1041019 generated at10/21/2017 16:20:36 needed for thread 1 ORA-00289: suggestion : /u/app/oracle/oradata/arch/1_8_957480185.dbf ORA-00280: change 1041019 for thread 1 isin sequence #8 ORA-00278: log file'/u/app/oracle/oradata/arch/1_7_957480185.dbf' no longer needed for thisrecovery ORA-00308: cannot open archived log'/u/app/oracle/oradata/arch/1_8_957480185.dbf' ORA-27037: unable to obtain file status --缺少序号为8的归档,这是因为redo还存在归档未切换 Linux-x86_64 Error: 2: No such file ordirectory Additional information: 3 ORA-01547: warning: RECOVER succeeded butOPEN RESETLOGS would get error below ORA-01194: file 1 needs more recovery to beconsistent ORA-01110: data file 1:'/u/app/oracle/oradata/racdg/system01.dbf' SQL> selecta.group#,a.sequence#,a.archived,a.status,b.member from v$log a,v$logfile bwhere a.group#=b.group#; GROUP# SEQUENCE# ARCHIV STATUS MEMBER ---------- ---------- ------------------------------------------------------------------------------------------------ 1 4 YES INACTIVE /u/app/oracle/oradata/racdg/redo01.log 3 6NO CURRENT /u/app/oracle/oradata/racdg/redo03.log --当前未归档的redo文件 2 5 YES INACTIVE /u/app/oracle/oradata/racdg/redo02.log SQL> recover database using backupcontrolfile; --再次执行,应用未归档的redo文件 ORA-00279: change 1041019 generated at10/21/2017 16:20:36 needed for thread 1 ORA-00289: suggestion :/u/app/oracle/oradata/arch/1_8_957480185.dbf ORA-00280: change 1041019 for thread 1 isin sequence #8 --这里看到未归档的redo文件,归档序号为6,而这里需要恢复序列为8的归档,所以应该是应用/u/app/oracle/oradata/racdg/redo02.log这个redo文件。 Specify log: {<RET>=suggested |filename | AUTO | CANCEL} /u/app/oracle/oradata/racdg/redo02.log Log applied. Media recovery complete. SQL> alter database open resetlogs; --以resetlogs的方式打开数据库 Database altered.
resetlogs与noresetlogs的区别说明:
norestlogs控制文件的scn是来自当前日志的highscn,而resetlogs控制文件的scn是来自数据文件。
noresetlogs会继续使用已经存在,且有效的logfiles,而resetlogs会初始化logs,重置log sequence号,创建一个新的incarnation
2.3.3通过rman备份控制文件进行不完全恢复思路
A. 若在线损坏,shutdown abort关闭数据库;若关闭数据库后损坏,到第二步
B. startup nomount启动数据库到nomount状态。
C. 使用rman从备份中恢复控制文件:restorecontrolfile from '/path';
D. alter database mount;启动数据库到mount状态
E. recover database using backupcontrolfile until cancel; 应用尽可能多的归档文件
G. 若丢失redo文件:
shutdown immediate并启动到startupnomount状态
alter database backup controlfile to traceas ‘/u/app/oracle/zwdir/controlfile.sql’; 生成创建控制文件的脚本
使用resetlogs方式创建控制文件
若未归档的redo文件可用:直接recover database,选用未归档的redo应用,alter database open resetlogs方式打开数据库
若未归档的redo文件不可用:设置隐含参数_allow_resetlogs_corruption=true跳过一致性检查,alterdatabase open resetlogs方式打开数据库
F. 若丢失归档文件:
shutdown immediate并启动到startupnomount状态
alter database backup controlfile to traceas ‘/u/app/oracle/zwdir/controlfile.sql’; 生成创建控制文件的脚本
使用noresetlogs方式创建控制文件
recover database using backup controlfileuntil cancel; 应用redo文件
alter database open resetlogs; 以resetlogs的方式打开数据库
若备份控制文件之后,数据库结构发生了变化,如新增了表空间或数据文件,需要进行不完全恢复
2.3.4通过rman备份控制文件进行不完全恢复演示
这里模拟演示模拟备份控制文件之后,新增表空间,归档文件全部丢失的不完全恢复的情况,丢失redo的情况,请看下一节“oracle数据库恢复系列之redo文件恢复”
[oracle@iscsi-asmzwdir]$ rman target / RMAN> backup current controlfile format'/u/app/oracle/zwdir/use_this_%U.ctl' tag='ctl'; --备份当前的控制文件 Starting backup at 2017/10/29 17:03:31 using target database control file insteadof recovery catalog piecehandle=/u/app/oracle/zwdir/use_this_08si88j3_1_1.ctl tag=CTL comment=NONE channel ORA_DISK_1: backup set complete,elapsed time: 00:00:01 Finished backup at 2017/10/29 17:03:33 [oracle@iscsi-asm racdg]$ sqlplus / assysdba SQL> create tablespace t_del_archdatafile '/u/app/oracle/oradata/racdg/t_del_arch01.dbf' size 20m; --新增表空间 Tablespace created. SQL> create user t_arch identified bypassword default tablespace t_del_arch; --新建用户 User created. SQL> grant resource,connect to t_arch; --授权 Grant succeeded. SQL> alter system switch logfile; --切换归档日志 System altered. SQL> create table t_arch.t1 as select *from dba_objects where rownum<1000; --新建表,并插入数据 Table created. SQL> alter system switch logfile; --切换归档日志 System altered. SQL> select count(*) from t_arch.t1; COUNT(*) ---------- 999 SQL> insert into t_arch.t1 select * fromt_arch.t1; --新增数据 999 rows created. SQL> commit; Commit complete. SQL> alter system switch logfile; --切换日志 System altered. SQL> select count(*) from t_arch.t1; --记录当前的数据,用作恢复的数据对比 COUNT(*) ---------- 1998 SQL> archive log list; --当前归档的序号 Database log mode Archive Mode Automatic archival Enabled Archive destination /u/app/oracle/oradata/arch Oldest online log sequence 5 Next log sequence to archive 7 Current log sequence 7 SQL> shutdown abort; --关闭数据库 ORACLE instance shut down. SQL> !rm control01.ctl control02.ctl --模拟控制文件丢失 SQL> !mv/u/app/oracle/oradata/arch/*.dbf /u/app/oracle/oradata/arch/tmp --模拟归档文件全部丢失 SQL> startup nomount; ORACLE instance started. SQL> !rman target / Recovery Manager: Release 11.2.0.4.0 -Production on Sun Oct 29 17:19:33 2017 Copyright (c) 1982, 2011, Oracle and/or itsaffiliates. All rights reserved. connected to target database: RACDG (notmounted) RMAN> restore controlfile from'/u/app/oracle/zwdir/use_this_08si88j3_1_1.ctl'; --通过旧备份控制文件进行恢复 Starting restore at 2017/10/29 17:19:44 using target database control file insteadof recovery catalog allocated channel: ORA_DISK_1 channel ORA_DISK_1: SID=10 device type=DISK channel ORA_DISK_1: restoring control file channel ORA_DISK_1: restore complete,elapsed time: 00:00:01 output filename=/u/app/oracle/oradata/racdg/control01.ctl output filename=/u/app/oracle/oradata/racdg/control02.ctl Finished restore at 2017/10/29 17:19:46 RMAN> exit Recovery Manager complete. SQL> alter database mount; Database altered. SQL> alter database backup controlfileto trace as '/u/app/oracle/zwdir/controlfile.sql'; Database altered. --此时归档文件已经丢失了,无法通过应用归档文件进行恢复,所以通过trc新建控制文件的方式进行恢复。 SQL> shutdown abort; --关闭数据库 ORACLE instance shut down. SQL> STARTUP NOMOUNT --启动到nomount状态 ORACLE instance started. SQL> CREATE CONTROLFILE REUSE DATABASE"RACDG" NORESETLOGS ARCHIVELOG MAXLOGFILES 16 MAXLOGMEMBERS 3 MAXDATAFILES 100 MAXINSTANCES 8 MAXLOGHISTORY 292 LOGFILE GROUP 1 '/u/app/oracle/oradata/racdg/redo01.log' SIZE 50M BLOCKSIZE 512, GROUP 2 '/u/app/oracle/oradata/racdg/redo02.log' SIZE 50M BLOCKSIZE 512, GROUP 3 '/u/app/oracle/oradata/racdg/redo03.log' SIZE 50M BLOCKSIZE 512 -- STANDBY LOGFILE DATAFILE '/u/app/oracle/oradata/racdg/system01.dbf', '/u/app/oracle/oradata/racdg/sysaux01.dbf', '/u/app/oracle/oradata/racdg/undotbs01.dbf', '/u/app/oracle/oradata/racdg/users01.dbf' CHARACTER SET WE8MSWIN1252 ; --以上新建控制文件的内容就是通过/u/app/oracle/zwdir/controlfile.sql里面的记录来进行新建的,里面共有两种方式,一种是noresetlogs,另一种是resetlogs的方式。这里因为redo文件是没有丢失的,所以使用nosesetlogs的方式。 Control file created. SQL> selectb.sequence#,a.member,b.status from v$logfile a, v$log b wherea.group#=b.group#; --查询redo信息 SEQUENCE# MEMBER STATUS ------------------------------------------------------- -------------------------------- 7 /u/app/oracle/oradata/racdg/redo01.log CURRENT 5 /u/app/oracle/oradata/racdg/redo02.log INACTIVE 6 /u/app/oracle/oradata/racdg/redo03.log INACTIVE SQL> select file#,name,status fromv$datafile; --查询数据文件信息,此时数据文件都是需要恢复 FILE# NAME STATUS ------------------------------------------------------- -------------- 1 /u/app/oracle/oradata/racdg/system01.dbf SYSTEM 2 /u/app/oracle/oradata/racdg/sysaux01.dbf RECOVER 3 /u/app/oracle/oradata/racdg/undotbs01.dbf RECOVER 4 /u/app/oracle/oradata/racdg/users01.dbf RECOVER SQL> recover database using backupcontrolfile until cancel; --恢复数据库,应用redo日志 ORA-00279: change 932267 generated at10/29/2017 17:07:56 needed for thread 1 ORA-00289: suggestion :/u/app/oracle/oradata/arch/1_6_958668721.dbf ORA-00280: change 932267 for thread 1 is insequence #6 Specify log: {<RET>=suggested |filename | AUTO | CANCEL} /u/app/oracle/oradata/racdg/redo03.log ORA-00279: change 932300 generated at10/29/2017 17:09:10 needed for thread 1 ORA-00289: suggestion :/u/app/oracle/oradata/arch/1_7_958668721.dbf ORA-00280: change 932300 for thread 1 is insequence #7 ORA-00278: log file'/u/app/oracle/oradata/racdg/redo03.log' no longer needed for this recovery Specify log: {<RET>=suggested |filename | AUTO | CANCEL} /u/app/oracle/oradata/racdg/redo01.log Log applied. Media recovery complete. SQL> alter database open resetlogs; --以resetlogs的方式打开数据库 Database altered. SQL> select file#,name,status fromv$datafile; --查看数据文件的状态,备份控制文件之后的表空间对应的数据文件需要进一步恢复。 FILE# NAME STATUS -------------------------------------------------------------- -------------- 1 /u/app/oracle/oradata/racdg/system01.dbf SYSTEM 2 /u/app/oracle/oradata/racdg/sysaux01.dbf ONLINE 3 /u/app/oracle/oradata/racdg/undotbs01.dbf ONLINE 4 /u/app/oracle/oradata/racdg/users01.dbf ONLINE 5/u/app/oracle/product/11.2.0/db_1/dbs/MISSING00005RECOVER SQL> alter database datafile 5 offline; --先将数据文件offline Database altered. SQL>alter database rename file'/u/app/oracle/product/11.2.0/db_1/dbs/MISSING00005' to'/u/app/oracle/oradata/racdg/t_del_arch01.dbf'; --重命名数据文件 Database altered. SQL> alter tablespace t_del_arch online; ---尝试将表空间online,报错 alter tablespace t_del_arch online * ERROR at line 1: ORA-01190: control file or data file 5 isfrom before the last RESETLOGS ORA-01110: data file 5: '/u/app/oracle/oradata/racdg/t_del_arch01.dbf' SQL> alter system set"_allow_resetlogs_corruption"=true scope=spfile; --修改参数,忽略SCN一致性校验 System altered. SQL> startup force; --再将数据库重新启动 ORACLE instance started. Total System Global Area 830930944 bytes Fixed Size 2257800 bytes Variable Size 536874104 bytes Database Buffers 285212672 bytes Redo Buffers 6586368 bytes Database mounted. Database opened. SQL> selectts#,file#,name,status,checkpoint_change# from v$datafile; --查询数据文件的情况 TS# FILE# NAME STATUS CHECKPOINT_CHANGE# ---------- ---------- ----------------------------------------------------------- ------------------ 0 1 /u/app/oracle/oradata/racdg/system01.dbf SYSTEM 953040 1 2 /u/app/oracle/oradata/racdg/sysaux01.dbf ONLINE 953040 2 3 /u/app/oracle/oradata/racdg/undotbs01.dbf ONLINE 953040 4 4 /u/app/oracle/oradata/racdg/users01.dbf ONLINE 953040 6 5/u/app/oracle/oradata/racdg/t_del_arch01.dbf RECOVER 0 SQL> alter session set events 'immediatetrace name adjust_scn level 1'; --通过设置adjust_scn使数据文件的CHECKPOINT_CHANGE#一致 Session altered. SQL> shutdown immediate; Database closed. Database dismounted. ORACLE instance shut down. SQL> startup mount; --启动到mount状态 ORACLE instance started. Total System Global Area 830930944 bytes Fixed Size 2257800 bytes Variable Size 536874104 bytes Database Buffers 285212672 bytes Redo Buffers 6586368 bytes Database mounted. SQL> recover until cancel; --恢复数据库,注意这里不是usingcontrolfile的方式 Media recovery complete. SQL> alter database datafile 2,3,4,5online; ---将所有的数据文件online,记得千万不用漏了 Database altered. SQL> alter database open resetlogs; --以resetlogs的方式打开数据库 Database altered. SQL> selectts#,file#,name,status,checkpoint_change# from v$datafile; --查询数据文件的情况 TS# FILE#NAME STATUS CHECKPOINT_CHANGE# ---------- ------------------------------------------------------- -------------- ------------------ 0 1 /u/app/oracle/oradata/racdg/system01.dbf SYSTEM 953040 1 2 /u/app/oracle/oradata/racdg/sysaux01.dbf ONLINE 953040 2 3/u/app/oracle/oradata/racdg/undotbs01.dbf ONLINE 953040 4 4 /u/app/oracle/oradata/racdg/users01.dbf ONLINE 953040 6 5 /u/app/oracle/oradata/racdg/t_del_arch01.dbf ONLINE 953040 SQL> select count(*) from t_arch.t1; --对比数据,发现存在数据丢失,出现这种情况,只能把损失降到最小 COUNT(*) ---------- 999 SQL> alter tablespace temp add tempfile'/u/app/oracle/oradata/racdg/temp01.dbf' reuse; --重建temp表空间 Database altered.
2.3.4通过trace备份控制文件进行完全恢复思路
A.备份控制文件到trace:alterdatabase backup controlfile to trace as '/path/controlfile.sql';
B.若在线损坏,直接alterdatabase backup controlfile to 'file';之后重启数据库
C.若关闭数据库后损坏,shutdownabort关闭数据库后,startup nomount启动数据库到nomount状态
D.使用备份到trace里创建控制文件的语句,以noresetlogs方式创建控制文件
E.recover database恢复数据库,恢复完后通过alterdatabase open打开数据库
F.重建temp表空间:altertablespace temp add tempfile '/path/temp01.dbf' reuse;
2.3.5通过trace备份控制文件进行完全恢复演示
oracle数据库提供了多种对控制文件的备份方式:rman、alterdatabase backup controlfile to trace、alter databasebackup controlfile to filename 这里只演示在线损坏时通过alter database backup controlfile to filename 的恢复,因为通过trace备份的恢复演示在“2.3.4通过rman备份控制文件进行不完全恢复演示”章节已经写了。
SQL> select open_mode from v$database; OPEN_MODE ---------------------------------------- READ WRITE SQL> !rm control01.ctl control02.ctl --这种情况主要是数据库还在运行状态,手贱误删的恢复 SQL> alter database backup controlfileto '/u/app/oracle/oradata/racdg/control01.ctl.bak'; --备份成二进制文件 Database altered. SQL> shutdown abort; ORACLE instance shut down. SQL> !cp control01.ctl.bak control01.ctl–直接拷贝使用 SQL> !cp control01.ctl.bak control02.ctl SQL> startup mount; --启动到mount状态 ORACLE instance started. Total System Global Area 830930944 bytes Fixed Size 2257800 bytes Variable Size 536874104 bytes Database Buffers 285212672 bytes Redo Buffers 6586368 bytes Database mounted. SQL> selecta.group#,a.sequence#,a.archived,a.status,b.member from v$log a,v$logfile bwhere a.group#=b.group#; GROUP# SEQUENCE# ARCHIV STATUS MEMBER ---------- ---------- --------------------- ---------------------------------------- 1 1 NO CURRENT /u/app/oracle/oradata/racdg/redo01.log 3 0 YES UNUSED /u/app/oracle/oradata/racdg/redo03.log 2 0 YES UNUSED /u/app/oracle/oradata/racdg/redo02.log SQL> recover database using backupcontrolfile until cancel; ORA-00279: change 932699 generated at10/31/2017 19:25:50 needed for thread 1 ORA-00289: suggestion :/u/app/oracle/oradata/arch/1_1_958850748.dbf ORA-00280: change 932699 for thread 1 is insequence #1 Specify log: {<RET>=suggested |filename | AUTO | CANCEL} /u/app/oracle/oradata/racdg/redo01.log –应用redo文件 Log applied. Media recovery complete. SQL> alter database open resetlogs; --以resetlogs的方式开启数据库 Database altered.
2.3.6通过trace备份控制文件进行不完全恢复思路
A.备份控制文件到trace:alterdatabase backup controlfile to trace as '/path/controlfile.sql';
B.shutdown abort关闭数据库后,startupnomount启动数据库到nomount状态
C. 若丢失redo文件,归档未丢失:
使用备份到trace里创建控制文件的语句,以resetlogs方式创建控制文件
alter database mount;启动数据库mount状态
若未归档的redo文件可用:直接recover database,选用未归档的redo应用,alter database open resetlogs方式打开数据库
若未归档的redo文件不可用:设置隐含参数_allow_resetlogs_corruption=true跳过一致性检查,alterdatabase open resetlogs方式打开数据库
D. 若丢失归档文件,redo未丢失:
使用备份到trace里创建控制文件的语句,以resetlogs方式创建控制文件
alter database mount;启动数据库mount状态
recover database using backup controlfileuntil cancel; 应用redo文件
alter database open resetlogs; 以resetlogs的方式打开数据库
若备份控制文件之后,数据库结构发生了变化,如新增了表空间或数据文件,需要进行不完全恢复
E. 若归档文件和redo文件都丢失:
步骤与若丢失归档文件,redo未丢失的恢复步骤一样,主要是以resetlogs方式创建控制文件
F.重建temp表空间:altertablespace temp add tempfile '/path/temp01.dbf' reuse;
2.3.7通过trace备份控制文件进行不完全恢复演示
请参照“2.3.4通过rman备份控制文件进行不完全恢复演示”,主要注意以下几点:
A.trace备份的控制文件并不是实时的备份,可能不是最新的,备份之后数据库结构可能发生了变化
B.对于此类恢复一定需要谨慎,恢复完之后马上做个全备,尽量将数据丢失降低到最小。
2.4所有的控制文件损坏,无备份场景
2.4.1通过手工重建控制文件进行恢复(noresetlogs)思路
A.shutdown abort关闭数据库后,startupnomount启动数据库到nomount状态
B.手工构造控制文件,以noresetlogs方式
C. alter database mount;启动数据库到mount状态
D. recover database using backupcontrolfile until cancel; 应用归档文件
E. 再执行recover databaseusing backup controlfile; 应用未归档的redo文件
F. alter database open resetlogs; 以resetlogs的方式打开数据库
2.4.2通过手工重建控制文件进行恢复(resetlogs)思路
A.shutdown abort关闭数据库后,startupnomount启动数据库到nomount状态
B.手工构造控制文件,以resetlogs方式
C. alter database mount;启动数据库到mount状态
D. recover database using backupcontrolfile until cancel; 应用归档文件(不考虑归档文件丢失)
E. 再执行recover databaseusing backup controlfile; 应用未归档的redo文件(不考虑redo文件丢失)
F. alter database open resetlogs; 以resetlogs的方式打开数据库
2.4.3通过手工重建控制文件进行恢复(resetlogs)演示
这里不作非常具体的演示了,前面的章节已经演示过了,就说明一下需要注意的几点:
因为是手动构建,所以在构建的过程中一定要仔细检查数据库清除数据库有多少文件,构建脚本如下:
CREATE CONTROLFILE REUSE DATABASE"RACDG" RESETLOGS ARCHIVELOG MAXLOGFILES 16 MAXLOGMEMBERS 3 MAXDATAFILES 100 MAXINSTANCES 8 MAXLOGHISTORY 292 LOGFILE GROUP 1 '/u/app/oracle/oradata/racdg/redo01.log' SIZE 50M BLOCKSIZE 512, GROUP 2 '/u/app/oracle/oradata/racdg/redo02.log' SIZE 50M BLOCKSIZE 512, GROUP 3 '/u/app/oracle/oradata/racdg/redo03.log' SIZE 50M BLOCKSIZE 512 -- STANDBY LOGFILE DATAFILE '/u/app/oracle/oradata/racdg/system01.dbf', '/u/app/oracle/oradata/racdg/sysaux01.dbf', '/u/app/oracle/oradata/racdg/undotbs01.dbf', '/u/app/oracle/oradata/racdg/users01.dbf', '/u/app/oracle/oradata/racdg/t_del_arch01.dbf' CHARACTER SET WE8MSWIN1252;
一定要都存放文件的路径下去ls –ltr列一下有哪些文件,千万不要把某个文件漏了。
B. 手动构建控制文件是存在一定的数据丢失风险的,一定得细心,细心,再细心。
2.4.4通过SNAPSHOTCONTROLFILE文件进行恢复思路
A.shutdown abort关闭数据库后,startupnomount启动数据库到nomount状态
B.直接拷贝snapshot controlfile成control file或通过rman恢复:restore controlfile from '$ORACLE_HOME/dbs/snapcf_@.f';
C. alter database mount;启动数据库到mount状态
D. recover database using backupcontrolfile until cancel; 应用归档文件(不考虑归档文件丢失)
E. 再执行recover databaseusing backup controlfile; 应用未归档的redo文件(不考虑redo文件丢失)
F. alter database open resetlogs; 以resetlogs的方式打开数据库
2.4.5通过SNAPSHOTCONTROLFILE文件进行恢复演示
RMAN> show snapshot controlfile name; --查看快照控制文件配置 using target database control file insteadof recovery catalog RMAN configuration parameters for databasewith db_unique_name RACDG are: CONFIGURE SNAPSHOT CONTROLFILE NAME TO'/u/app/oracle/product/11.2.0/db_1/dbs/snapcf_racdg.f'; # default --如果是RAC,一般建议将该配置文件配置在ASM或ACFS等共享存储上。 SQL> !rm control01.ctl control02.ctl --模拟控制文件丢失 SQL> shutdown abort; --强制关闭数据库 ORACLE instance shut down. SQL> startup nomount; --启动数据库到mount状态 ORACLE instance started. SQL> !cp/u/app/oracle/product/11.2.0/db_1/dbs/snapcf_racdg.f control01.ctl --直接拷贝快照控制文件 SQL> !cp/u/app/oracle/product/11.2.0/db_1/dbs/snapcf_racdg.f control02.ctl SQL> !ls -ltr control*.ctl -rw-r----- 1 oracle oinstall 9748480Nov 1 18:48 control01.ctl -rw-r----- 1 oracle oinstall 9748480Nov 1 18:48 control02.ctl --或者使用RMAN进行恢复:restorecontrolfile from ‘/u/app/oracle/product/11.2.0/db_1/dbs/snapcf_racdg.f’; SQL> alter database mount; --启动数据库到mount状态 Database altered. SQL> recover database using backupcontrolfile until cancel; --应该归档文件 ORA-00279: change 948557 generated at10/31/2017 22:00:44 needed for thread 1 ORA-00289: suggestion :/u/app/oracle/oradata/arch/1_2_958851405.dbf ORA-00280: change 948557 for thread 1 is insequence #2 Specify log: {<RET>=suggested |filename | AUTO | CANCEL} auto ORA-00308: cannot open archived log'/u/app/oracle/oradata/arch/1_4_958851405.dbf' ORA-27037: unable to obtain file status Linux-x86_64 Error: 2: No such file or directory Additional information: 3 ORA-00308: cannot open archived log'/u/app/oracle/oradata/arch/1_4_958851405.dbf' ORA-27037: unable to obtain file status Linux-x86_64 Error: 2: No such file ordirectory Additional information: 3 ORA-01547: warning: RECOVER succeeded butOPEN RESETLOGS would get error below ORA-01194: file 1 needs more recovery to beconsistent ORA-01110: data file 1:'/u/app/oracle/oradata/racdg/system01.dbf' SQL> selecta.group#,a.sequence#,a.archived,a.status,b.member from v$log a,v$logfile bwhere a.group#=b.group#; GROUP# SEQUENCE# ARCHIV STATUS MEMBER ---------- ---------- -------------------------- ------------------------------------------------ 1 1 YES INACTIVE /u/app/oracle/oradata/racdg/redo01.log 3 0 YES UNUSED /u/app/oracle/oradata/racdg/redo03.log 2 2 NO CURRENT /u/app/oracle/oradata/racdg/redo02.log SQL> recover database using backupcontrolfile until cancel; --应用redo文件 ORA-00279: change 948625 generated at11/01/2017 18:41:13 needed for thread 1 ORA-00289: suggestion :/u/app/oracle/oradata/arch/1_4_958851405.dbf ORA-00280: change 948625 for thread 1 is insequence #4 Specify log: {<RET>=suggested |filename | AUTO | CANCEL} /u/app/oracle/oradata/racdg/redo01.log Log applied. Media recovery complete. SQL> alter database open resetlogs; --以resetlogs的方式开启数据库 Database altered.
总结:
1、从上可以说是涉及到各个场景控制文件的损坏恢复,相对是比较全面的,但生产环境下的恢复情况要比我们预想的复杂的多,应根据各种情况,尽量将损失降低到最小。
2、还有个在linux环境未关库情况下,造成控制文件误删,可通过/proc进程号恢复,这边就不做介绍了。
3、最后得强调一下,有效的备份重于一切,对生产环境存敬畏之心,千万不要存侥幸心理,墨菲定律就在我们身边。