【赛迪网－it技术报道】一、安装

环境：

primary 数据库：sun01

standby 数据库：sun02

oracle_sid:bnet

1. 拷贝$oracle_base 到新机器上

2. 让primary数据库处于归档模式，对primary数据库做一个rman热备份

startup mount

alter database archivelog;

archive log list;

rman target sys/oracle

backup database format "/....";

3.创建data guard数据库控制文件

alter database create standby controlfile as '/....ctl';

4.设置监听

primary：

bnetguard =

( (description =

(address_list =

(address = (protocol = tcp)(host = sun02)(port = 1522))

)

(connect_data =

(server = dedicated)

(service_name = bnet)

)

)

standby:

listener.ora

listener =

(description_list =

(description =

(address_list =

(address = (protocol = ipc)(key = extproc))

)

(address_list =

(address = (protocol = tcp)(host = sun01)(port = 1522))

)

)

)

tnsname.ora

bnet =

( (description =

(address_list =

(address = (protocol = tcp)(host = sun02)(port = 1522))

)

(connect_data =

(server = dedicated)

(service_name = bnet)

)

)

5. 设置主机参数

alter database force logging;

主库强制归档

主机参数：

fal_client=bnetguard

fal_server=bnet

standby_file_management=auto

remote_login_passwordfile=exclusive

备机参数：

db_file_name_convert

log_file_name_convert

standby_file_management='auto'

standby_archive_dest=/export/home/oracle/oradata/bnet/arch

6.创建standby 数据库

orapwd

sqlplus "sys/oracle as sysdba"

startup nomount pfile='''

alter database mount standby database;

exit

rman target sys/oracle nocatalog

restore database

7.设置自动归档

alter system set log_archive_dest_2='server=bnetguard reopen=200'

8.启动data guard自动恢复状态

sqlplus "sys/oracle as sysdba"<

recover managed standby database;

exit

eof

9.测试

alter system archive log current;

二、维护：

使standby 数据库处与 read-only状态

reover cancel/recover managed standby database cancel;

取消恢复模式

alter database open read only;

启动数据库到只读状态

recover standby database;

使数据库切换到恢复模式

recover managed standby database time out 15;

如果超过15分钟没有传日志

restarting dead background process qmn0

"restarting dead background process qmn" message can continually

appear in the alert log during startup in rac environment.

workaround:

set aq_tm_processes=0 in the init.ora.

it can be set to desired value through

"alter system set aq_tm_processes=n" once the database is open completely.

alter database activate standby database;