ORACLE 11g 用Duplicate恢复Data Guard 备库详细过程

Posted 2020-08-20

 

1、先查找备库控制文件路径

先在备库上找出控制文件的路径，通过和主库一样，不过为了以防万一，还是check为好。

SQL>  select name from v$controlfile;                   NAME -------------------------------------------------------------------------------- /Oracle/app/oracle/oradata/powerdes/control01.ctl /oracle/app/oracle/fast_recovery/powerdes/control02.ctl   SQL>

 

 

 

2、准备控制文件备份文件

 

在主库上备份当前的控制文件，记得生成的备份文件要在执行全备之前生成，而且命令行也是带for standby字样的backup current controlfile for standby format‘/home/oracle/ctlfile.bak‘才行，也可以手动提前生成。

RMAN> backup current controlfile for standby format ‘/home/oracle/ctlfile.bak‘;   Starting backup at 2016-11-23 21:20:41 using channel ORA_DISK_1 channel ORA_DISK_1: starting full datafile backup set channel ORA_DISK_1: specifying datafile(s) in backup set including standby control file in backup set channel ORA_DISK_1: starting piece 1 at 2016-11-23 21:20:42 channel ORA_DISK_1: finished piece 1 at 2016-11-23 21:20:43 piece handle=/home/oracle/ctlfile.bak tag=TAG20161123T212041 comment=NONE channel ORA_DISK_1: backup set complete, elapsed time: 00:00:01 Finished backup at 2016-11-23 21:20:43   RMAN>

 

 

然后等备份任务结束后，在主库上，将备份前生成的控制文件、最近的一份备份文件到备库的相同目录：

scp /oracle/app/oracle/oradata/powerdes/control01.ctl 192.168.3.112:/home/oracle/ scp -r /data/backup/data/2016-11-23 192.168.3.112:/home/oracle/

 

 

 

3、检查主库备库环境

获取备库的tns：（我的是earth_m2）

cat $ORACLE_HOME/network/admin/tnsnames.ora

earth_m2 =   (DESCRIPTION =     (ADDRESS_LIST =       (ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.3.112)(PORT = 1521))     )     (CONNECT_DATA =       (SERVICE_NAME = powerdes)     )   )

 

 

（1）查看主库备库SID名字是否统一，这个很重要

# 备库： [[email protected]t1_dbm1_3_112 data]$  echo $ORACLE_SID powerdes [[email protected]t1_dbm1_3_112 data]$   # 主库： [[email protected]t1_dbm1_3_111 ~]$  echo $ORACLE_SID powerdes [[email protected]t1_dbm1_3_111 ~]$

 

（2）查看主库备份传输密码是否一致：

# 备库： [[email protected]t1_dbm1_3_112 dbs]$ cd $ORACLE_HOME/dbs [[email protected]t1_dbm1_3_112 dbs]$  strings orapwpowerdes ]\[Z ORACLE Remote Password file INTERNAL 89F7167638478F13 0Cv# B6E56CBCA47429E2 [[email protected]t1_dbm1_3_112 dbs]$   # 主库： [[email protected]t1_dbm1_3_111 ~]$ cd $ORACLE_HOME/dbs [[email protected]t1_dbm1_3_111 dbs]$ strings orapwpowerdes ]\[Z ORACLE Remote Password file INTERNAL 89F7167638478F13 0Cv# B6E56CBCA47429E2 [[email protected]t1_dbm1_3_111 dbs]$

 

 

4、启动备库到nomount状态

在备库上，直接copy覆盖控制文件

cp /home/oracle/ctlfile.bak /oracle/app/oracle/oradata/powerdes/control01.ctl cp /home/oracle/ctlfile.bak /oracle/app/oracle/fast_recovery/powerdes/control02.ctl

 

然后启动到nomount状态

SQL> startup nomount ORACLE instance started.   Total System Global Area 8551575552 bytes Fixed Size              2270360 bytes Variable Size              7247760232 bytes Database Buffers    1291845632 bytes Redo Buffers                 9699328 bytes SQL>

 

 

 

 

5、使用duplicate恢复备库

在主库上，使用rman登录远程备库earth_m2

[[email protected]t1_dbm1_3_111 ~]$ rlwrap rman target / auxiliary sys/04181123@earth_m2   Recovery Manager: Release 11.2.0.4.0 - Production on Wed Nov 23 20:05:54 2016   Copyright (c) 1982, 2011, Oracle and/or its affiliates.  All rights reserved.   connected to target database: POWERDES (DBID=3481287162) connected to auxiliary database: POWERDES (not mounted)   RMAN>

 

 

在rman命令行里面执行恢复命令：

run {
allocate auxiliary channel c1 devicetype disk;
allocate auxiliary channel c2 devicetype disk;
duplicate target database for standbynofilenamecheck dorecover;
release channel c1;
release channel c2;
}
#nofilenamecheck：必须指定NOFILENAMECHECK参数，避免覆盖primary数据库的当前的数据文件。另外主从路径一致，不需要执行文件路径以及check了
#release ：这是关闭 前两行开启的通道

 

执行过程比较漫长，屏幕滚动信息比较多，这里不贴出来了，省略过……

 

执行命令后，需要很长时间，需要等待，因为需要写入文件到备库的数据文件目录里面，找个时候，在备库的alert log后台有如下记录日志，表明正在写数据文件：

Full restore complete of datafile 4 /oracle/app/oracle/oradata/powerdes/users01.dbf.  Elapsed time: 0:00:00   checkpoint is 32945676   last deallocation scn is 3 Wed Nov 23 20:14:42 2016 Full restore complete of datafile 3 /oracle/app/oracle/oradata/powerdes/undotbs01.dbf.  Elapsed time: 0:00:05   checkpoint is 32945676   last deallocation scn is 32897307   Undo Optimization current scn is 32934775 Wed Nov 23 20:15:50 2016 Full restore complete of datafile 5 /oracle/app/oracle/oradata/powerdes/example01.dbf.  Elapsed time: 0:01:18   checkpoint is 32945676   last deallocation scn is 32806636 …………

 

 

6、ORA-01152:问题

最后rman命令行里面duplicate命令执行完，窗口界面上有如下报错信息:

Signalling error 1152 for datafile 1! Errors in file /oracle/app/oracle/diag/rdbms/earth_m2/powerdes/trace/powerdes_pr00_4069.trc: ORA-01547: warning: RECOVER succeeded but OPEN RESETLOGS would get error below ORA-01152: file 1 was not restored from a sufficiently old backup ORA-01110: data file 1: ‘/oracle/app/oracle/oradata/powerdes/system01.dbf‘ ORA-1547 signalled during: alter database recover cancel... ^C

 

有报错没有关系，接下来直接开始应用日志，启动归档日志传输，通过应用归档日志来保证主备数据的一致性：

# 先开始启动传输开始应用日志 SQL> alter database recover managed standby database using current logfile disconnect from session;   Database altered.   SQL># 实时观察应用日志状态 SQL> select sequence#,applied from v$archived_log order by sequence# asc; ……  SEQUENCE# APPLIED ---------- ---------       2334 YES       2335 YES   2301 rows selected.   SQL> # 等观察应用日志结束后，退出日志传输 SQL> alter database recover managed standby database cancel;      Database altered.   SQL> # 将备库从mount启动到open状态 SQL> alter database open;              Database altered.   SQL> # 查看备库状态，就会变成READ ONLY SQL> select database_role,open_mode from v$database;      DATABASE_ROLE     OPEN_MODE ---------------- -------------------- PHYSICAL STANDBY READ ONLY   SQL> # 然后再次启动日志应用 SQL> alter database recover managed standby database using current logfile disconnect from session;   Database altered.   SQL>  # 查看open_mode，有了READ ONLY WITH APPLY字样，OK SQL> select database_role,open_mode from v$database;     DATABASE_ROLE     OPEN_MODE ---------------- -------------------- PHYSICAL STANDBY READ ONLY WITH APPLY   SQL>

 

 

 

至此，Data Guard备库修复完成，如果遇到了ORA-01152问题，不需要慌张，可以试试应用日志来恢复数据。