第11步：数据库冗余

目前在ASM上控制文件只有一份，需要给控制文件做镜像

1. 1. 查看当前控制文件

代码1

[root@sgdb1 ~]# su - oracle

[oracle@sgdb1 ~]$ export ORACLE_SID=sgdb1 //第2个节点为：export ORACLE_SID=sgdb2

[oracle@sgdb1 ~]$ sqlplus / as sysdba

SQL> show parameter control_files

NAME TYPE VALUE

------------------------------------ ----------- ------------------------------

control_files string +BJ_SGDB_DATA/sgdb/controlfile/current.256.887810989

1. 1. 复制控制文件

（1）两个节点分别关闭数据库

代码2

[oracle@sgdb1 ~]$ sqlplus / as sysdba

SQL> shutdown immediate;

Database closed.

Database dismounted.

ORACLE instance shut down.

代码3

[oracle@sgdb2 ~]$ sqlplus / as sysdba

SQL> shutdown immediate;

Database closed.

Database dismounted.

ORACLE instance shut down.

（2）启动数据库到nomount状态（实例之一即可）

代码4

SQL> startup nomount;

ORACLE instance started.

Total System Global Area 1.0055E+10 bytes

Fixed Size 2261888 bytes

Variable Size 5435821184 bytes

Database Buffers 4596957184 bytes

Redo Buffers 19742720 bytes

查看控制文件位置

SQL> show parameter control_files;

NAME TYPE VALUE

------------------------------------ ----------- ------------------------------

control_files string +BJ_SGDB_DATA/sgdb/controlfile/current

.256.887810989

（3）使用RMAN复制控制文件（实例之一即可）

代码5

[oracle@sgdb1 ~]$ rman target /

RMAN> restore controlfile to '+BJ_SGDB_DATA' from

'+BJ_SGDB_DATA/sgdb/controlfile/current.256.887810989';

（4）查看新生成的控制文件副本

注意：新生成的文件名目前没有看到，可以通过grid用户，使用asmcmd查看

代码6

[grid@sgdb1 ~]$ export ORACLE_SID=+ASM1

[grid@sgdb1 ~]$ asmcmd

ASMCMD> cd DATA/SGDB/CONTROLFILE

ASMCMD> ls

Current.256.887810989

current.269.887816799

显然，新生成的文件名字为current.269.887816799

（5）修改SPFILE

注意RAC不能像单实例那样直接修改pfile，不要轻易使用“create pfile from spfile;”命令创建PFILE文件，因此还是老老实实地修改spfile吧：

代码7

[oracle@sgdb1 ~]$ sqlplus / as sysdba

当前状态应该已经为nomount

SQL> select status from v$instance;

STATUS

------------

STARTED

修改参数文件

SQL> alter system set control_files = '+BJ_SGDB_DATA/sgdb/controlfile/current.256.887810989', '+BJ_SGDB_DATA/sgdb/controlfile/current.269.887816799' scope = spfile;

System altered.

（6）分别重启两个数据库（sgdb1 & sgdb2），查看控制文件

代码8

SQL> startup force;

SQL> show parameter control_files;

NAME TYPE VALUE

------------------------------------ ----------- ------------------------------

control_files string +BJ_SGDB_DATA/sgdb/controlfile/current

.256.887810989, +BJ_SGDB_DATA/sgdb/con

trolfile/current.269.887816799

1. REDO LOG冗余

目前REDO LOG一共有8组，每个组只有一名成员，需要做镜像

代码9

[oracle@sgdb1 ~]$ sqlplus / as sysdba

拷贝下面语句，执行添加成员

alter database add logfile member '+BJ_SGDB_DATA' to group 1;

alter database add logfile member '+BJ_SGDB_DATA' to group 2;

alter database add logfile member '+BJ_SGDB_DATA' to group 3;

alter database add logfile member '+BJ_SGDB_DATA' to group 4;

alter database add logfile member '+BJ_SGDB_DATA' to group 5;

alter database add logfile member '+BJ_SGDB_DATA' to group 6;

alter database add logfile member '+BJ_SGDB_DATA' to group 7;

alter database add logfile member '+BJ_SGDB_DATA' to group 8;

终于，OracleRAC安装完成！

==============================================================================

1.rman-00554:initialization of internal recovery manager package failed

rman-04005:error from target database:

ORA-12162:TNS:net service name is incorrectly specified

解决过程：

忘记了，下次在遇到留意下。

2.启动RAC其中一个节点正常，启动第二个节点报错，提示如下：

SQL> startup force;

ORACLE instance started.

Total System Global Area 1068937216 bytes

Fixed Size 2260088 bytes

Variable Size 444597128 bytes

Database Buffers 616562688 bytes

Redo Buffers 5517312 bytes

ORA-01105: mount is incompatible with mounts by other instances

ORA-01606: parameter not identical to that of another mounted instance

解决过程：

之前的步骤修改SPFILE修改参数文件，应该两个节点同时关闭，同时以nomount方式启动，同时执行

alter system set control_files = '+BJ_SGDB_DATA/sgdb/controlfile/current.256.887810989', '+BJ_SGDB_DATA/sgdb/controlfile/current.269.887816799' scope = spfile;

System altered.

问题解决。