oracle11gR2 Dataguard的配置过程

以前在学校从来玩过集群,上班了公司要用户dataguard,叫我来配置,这个过程之纠结,做了些记录:

 

1.      首先是在主库上创建一个数据库,取名叫orcl,配置好这个数据库的ORACLE_SID,使用sqlplus进入到数据库,执行下面的命令,设置数据库为归档模式:

a)        SQL>shutdown immediate

b)        SQL>startup mount

c)        SQL>alter database archivelog;

d)        SQL>alter database open;

e)        SQL>archive log list;

 

2.      备库要成为主库的完全相同的复制,就必须接受来自主库的重做日志。这就要启用数据库的强制日志功能:

a)        SQL>alter database force logging;

b)        SQL>select name, force_logging from v$database;

3.      确认当主库添加或删除文件是,这些文件也会在备库添加或删除,启动的sql语句:

SQL> alter system setstandby_file_management = 'AUTO';

4.      接下来我们就要创建主库的备用日志文件,备库使用备份日志文件来保持从主库接受到的日志文件,当主库转换为备库后,也要使用日志文件,建立4组,一般比在线日志多一组,建立的方法:

a)        alterdatabase add standby logfile group 11 (‘/opt/oracle/app/oradata/orcl/g11m01.sdo’,’/opt/oracle/app/oradata/orcl/g11m02.sdo’)size 51M;

b)        alterdatabase add standby logfile group 12 (‘/opt/oracle/app/oradata/orcl/g12m01.sdo’,’/opt/oracle/app/oradata/orcl/g12m02.sdo’)size 51M;

c)        alterdatabase add standby logfile group 13 (‘/opt/oracle/app/oradata/orcl/g13m01.sdo’,’/opt/oracle/app/oradata/orcl/g13m02.sdo’)size 51M;

d)        alterdatabase add standby logfile group 14 (‘/opt/oracle/app/oradata/orcl/g14m01.sdo’,’/opt/oracle/app/oradata/orcl/g14m02.sdo’)size 51M;

 

5.      检查数据库的db_unique_name 参数是否设置,没有就设置:

a)        SQL>show paramter db_unique_name;

b)        SQL>alter system set db_unique_name=some_name scope=spfile;

 

 

6.      开启主库的闪回功能,若主库出现问题,我们就可以从闪回到故障前,我们就可以避免从备库中来恢复数据,启用闪回功能,必须先配置快速恢复区:

a)        SQL>alter system set db_recovery_file_dest='/opt/oracle/app/flash_recovery_area ';

b)        SQL>alter system set db_recovery_file_dest_size=400G;

7.      配置好回闪区后就要开启回闪日志的功能;

a)        SQL>alter database flashback on;

b)        SQL>select flashback_on from v$database;

 

8.      配置listener.ora和tnsnames.ora,

a)        在主库listener.ora中添加下面代码:

LISTENER =

 (DESCRIPTION_LIST =

   (DESCRIPTION =

     (ADDRESS = (PROTOCOL = TCP)(HOST = localhost)(PORT = 1521))

    )

  )

 

ADR_BASE_LISTENER = /opt/oracle/app

 

 

 

SID_LIST_LISTENER=

   (SID_LIST =

       (SID_DESC =

           (GLOBAL_DBNAME = orcl)

           (ORACLE_HOME = /opt/oracle/app/product)

           (SID_NAME = orcl)

        )

    )

b)        在备库的listener.ora中添加下面代码:

LISTENER =

 (DESCRIPTION_LIST =

   (DESCRIPTION =

     (ADDRESS = (PROTOCOL = TCP)(HOST = localhost)(PORT = 1521))

    )

  )

 

ADR_BASE_LISTENER = /opt/oracle/app/oracle

 

 

SID_LIST_LISTENER=

   (SID_LIST =

       (SID_DESC =

           (GLOBAL_DBNAME = orcl)

           (ORACLE_HOME = /opt/oracle/app/oracle/product/11.2.0/dbhome_1)

           (SID_NAME = orcl)

        )

    )

c)        在主库的tnsname.ora中添加下面的代码:(注释掉以前的内容)

cloudvastprimary =

 (DESCRIPTION =

   (ADDRESS = (PROTOCOL = TCP)(HOST = localhost)(PORT = 1521))

   (CONNECT_DATA =

     (SERVER = DEDICATED)

     (SERVICE_NAME = orcl)

    )

  )

 

cloudvaststandby =

 (DESCRIPTION =

   (ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.10.133)(PORT = 1521))

   (CONNECT_DATA =

     (SERVER = DEDICATED)

     (SERVICE_NAME = orcl)

    )

  )

d)        在备库的tnsname.ora中添加下面的代码:

cloudvaststandby =

 (DESCRIPTION =

   (ADDRESS = (PROTOCOL = TCP)(HOST = localhost)(PORT = 1521))

   (CONNECT_DATA =

     (SERVER = DEDICATED)

     (SERVICE_NAME = orcl)

   )

 )

 

cloudvastprimary =

 (DESCRIPTION =

   (ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.10.112)(PORT = 1521))

   (CONNECT_DATA =

     (SERVER = DEDICATED)

     (SERVICE_NAME = orcl)

   )

 )

9.      完成之后就可以使用tnsping在测试一下两台机器是否连通的

a)        在主机上tnsping cloudvaststandby

b)        在备机上tnsping cloudvastprimary

c)        若是没有通,出现了错误,那么就关掉防火墙

 

 

10.  配置重做日志的传输:

a)        配置归档位置:

SQL> alter system setlog_archive_dest_1 = 'location=use_db_recovery_file_dest valid_for=(all_logfiles,all_roles) db_unique_name=cloudvastprimary';

b)        配置重做日志传输到备库:

SQL> alter system set log_archive_dest_2= 'service=cloudvaststandby async valid_for=(online_logfile,primary_role)db_unique_name= cloudvaststandby;

c)        配置fal_server ,这个参数的作用就是当日志文件在传输出现了问题时,备库要到哪儿去到归档日志。比如在对备库进行维护期间,没有日志传输过来,这是重做日志间就会出现缺口,设置这个参数,备库就去找缺少的日志:

SQL> alter system setfal_server = 'cloudvaststandby’;

d)        让主库知道dataguard配置里的另外一个库的名字:

SQL> alter system setlog_archive_config = 'dg_config=(cloudvastprimary, cloudvaststandby)';

11.  备库环境的准备:

a)        在主库使用语句创建一个启动参数文件

SQL> create pfile fromspfile;

b)        然后就使用scp命令把主库的密码文件和刚才创建的启动文件传递到备库上,放到$ORACLE_HOME/dbs/这个目录下面,修改密码文件名为备库的SID,这里主备库都是orcl,所以就不修改

c)        打开启动参数文件initorcl.ora,修改里面参数的路径,改为备库的路径,修改db_unique_name=cloudvaststandby,   fal_server=cloudvastprimary,  log_archive_dest_1的db_unique_name修改为备库的唯一名,log_archive_dest_2修改对应的服务名和数据库的唯一名为cloudvastprimary

d)        若是主库和备库的安装路径不一样,还有设置dbfile_name_convert  log_file_name_convert

e)        根据配置文件创建相应的文件目录

f)         将备库的信息添加到/etc/oratab文件

g)        现在可以启动备库的实例来创建数据库了,启动后创建spfile

SQL> startup nomountpfile=initJED2.ora

SQL> create spfile from pfile;

SQL> shutdown

SQL> startup nomount

SQL> show parameter spfile

SQL> exit

在这里可能会有问题,我的主库系统是centos6.3的,而备库的系统是centos6.4的,所以在启动数据库时出现了错误:

                     SQL> startup nomountpfile=initorcl.ora

ORA-00845: MEMORY_TARGET notsupported on this system

解决的方法有两种:一种就是把MEMORY_TARGET设置小,另一种就是该打

/dev/shm的大小

 

12.  备库的创建,使用rman命令来创建备库,在主库上执行下面的命令:

RMAN> connect target sys@cloudvastprimary

RMAN> connect catalog system@cloudvastprimary

RMAN> connect auxiliary sys@cloudvaststandby

RMAN> duplicate target databasefor standby from active database;

 

这里可能会出现的错误:

a)       RMAN-04006: error from auxiliarydatabase: ORA-12528: TNS:listener: all appropriate instances are blocking newconnections

 

    仔细检查listener.ora,tnsname.ora文件,重启监听器

 

b)       Starting Duplicate Db at 27-APR-13

RMAN-00571: ===========================================================

RMAN-00569: ===============ERROR MESSAGE STACK FOLLOWS ===============

RMAN-00571:===========================================================

RMAN-03002: failure ofDuplicate Db command at 04/27/2013 16:00:07

RMAN-12010: automaticchannel allocation initialization failed

RMAN-06428: recoverycatalog is not installed

 

 

解决方法:

进入到sqlplus中,执行下面的语句:

SQL> create tablespacerman_ts datafile '/opt/oracle/app/oradata/orcl/rman_ts.dbf' size 50M;

SQL> grantrecovery_catalog_owner to system;

进入到rman中执行下面的语句:

create catalog tablespacerman_ts;

register database;

c)       当看到Finished Duplicate Db at 27-APR-13的时候就说明已经完成了备库的创建

 

13.  启用和取消重做日志应用

a)        SQL>ALTER DATABASE RECOVER MANAGED STANDBY DATABASE USING CURRENT LOGFILE FDISCONNECT FROM SESSION;

b)        ALTERDATABASE RECOVER MANAGED STANDBY DATABASE CANCEL;

14.  确认日志应用是否正常

SQL> select DEST_ID,STATUS, DESTINATION, ERROR from V$ARCHIVE_DEST where DEST_ID<=2;

目的地状态显示的是valid

15.  检查日志是否有缺口

SQL> select STATUS,GAP_STATUS from V$ARCHIVE_DEST_STATUS where DEST_ID = 2;

正常的应该返回valid和no gap

16.  若想要查询备库的数据,就打开数据库

SQL> ALTER DATABASE OPENREAD ONLY;

 

17.  到此dataguard的配置就完成了

 

 

 

 

你可能感兴趣的:(oracle11gR2,Oracle集群,dataguard配置)