Oracle高可用之dataguard
DataGuard是一种数据库级别的HA方案,最主要功能是冗灾、数据保护、故障恢复等。
在生产数据库的"事务一致性"时,使用生产库的物理全备份(或物理COPY)创建备库,备库会通过生产库传输过来的归档日志(或重做条目)自动维护备用数据库。将重做数据应用到备用库。本文介绍使用RMAN备份创建备库(dataguard).
一:.环境
1.软件环境
操作系统Red Linux Enterprise as 5
数据库版本Oracle 10g release 2
2.primary databae
IP:192.168.18.1
ORACLE_SID=db1
db_unique_name=db1
3.standby database
IP:192.168.18.2
ORACLE_SID=standby
db_unique_name=standby
二,主数据库(db1)做准备
1.设置主数据库为Force logging
SQL> alter database force logging;
|
2.创建密码文件
cd $ORACLE_HOME/dbs/
orapwd file=orapwdb1 password=123456 force=y
|
3.修改主库的初始化参数
alter system set log_archive_config='dg_config=(db1,standby)' scope=both;
alter system set log_archive_dest_1='location=/u01/db1/arch' scope=both;
alter system set db_unique_name='db1' scope=both;
|
5.生成数据库备份
RMAN> connect target sys/123456
RMAN> backup database format='/oracle/rmanback/%d_%s.dbf' plus archivelog;
[oracle@oracle rmanback]$ ls
DB1_1.dbf DB1_2.dbf
|
6.生成备库的control file
SQL>alter database create standby controlfile as '/oracle/rmanback/ctontrl01.ctl
|
7.配置listener.ora 和tnsnames.ora文件
启动lintener.ora,
[oracle@oracle dbs]$ lsnrctl status
LSNRCTL for Linux: Version 10.2.0.4.0 - Production on 14-JUN-2009 02:54:29
Service "db1" has 1 instance(s).
Instance "db1", status READY, has 1 handler(s) for this service...
Service "db1_XPT" has 1 instance(s).
Instance "db1", status READY, has 1 handler(s) for this service...
The command completed successfully
|
配置rnsnames.ora
vi $ORACLE_HOME/network/admin/tnsnames.ora
db1 =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.18.1)(PORT = 1521))
)
(CONNECT_DATA =
(SERVICE_NAME = db1)
standby =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.18.2)(PORT = 1521))
)
(CONNECT_DATA =
(SERVICE_NAME = standby)
)
)
|
三.创建standby database
1.设置环境变量并安装oracle软件
#
环境变量设成与主库一样
#
只安装软件不安装库
./runInstaller –silent –responseFile /tmp/installoracle.rsp
|
2.建立相关的目录
mkdir –p $ORACLE_BASE/admin/
cd $ORACLE_BASE/admin/ORACLE_SID
mkdir adump bdump cdump dpdump udump
|
3.建立密码文件
cd $ORACLE_HOME/dbs/
orapwd file=orapwSID password=123456 force=y
#
注要密码要与主库的一样
,
否则会归档失败
|
3.建立参数文件(pfile)
db_name = standby
shared_pool_size = 120M
undo_management = AUTO
undo_tablespace = undotbs
sga_max_size = 200M
sga_target = 160M
db_2k_cache_size = 4M
standby_file_management=AUTO
fal_server='db1'
fal_client='standby'
log_archive_dest_1='location=/u01/app/oracle/product/10.2.0/dbs/arch'
log_archive_dest_2='SERVICE=db1 REOPEN=300'
log_archive_dest_state_1='ENABLE'
log_archive_dest_state_2='ENABLE'
|
4.CP主数据库RMAN备份及控制文件到备库
注意:备份存放位置要与primary database RMAN备份文件的位置相同. 控制文件存放位置要与生成standby database controlfile的位置相同
scp /oracle/rmanback/*.dbf [email protected]:/oracle/rmanback/
|
5.利用备用的控制文件,把备用数据库启到mount
SQL>connect / as sysdba
connnpcted to an idle instance.
SQL>startup nomount pfile=$ORACLE_HOME/dbs/initstandby.ora
SQL>alter database mount standby database;
|
6. 配置listener.ora 和tnsnames.ora文件.
与主库相同启动listener,tnsnames.ora也与主库配置的一样,
当主备库的监听都启动后,进行测试,以例下面能顺利进行
tnsping db1
tnsping standby
SQL> sqlplus sys/123456@db1
SQL> sqlplus sys/123456@standby
|
7.转储数据库
RMAN>connect target /
connected to target database:TEST(DBID=788075692)
RMAN> restore database
|
8.恢复数据库。
SQL>recover managed standby database disconnect from session;
#如果有需要
应
用的日志并想手工
应
用,可以运行如下命
令
SQL>recover automatic standby database;
|
9.检查standby database是否创建成功
1.在primary database 上切换日志
SQL> alter system switch logfile
|
2.在primary database上运行下面的语句
SQL> select max(sequence#) from v$archived_log;
MAX(SEQUENCE#)
--------------
17
|
3. 在standby database上运行下面的语句
SQL> select sequence# ,applied from v$archived_log order by sequence#;
SEQUENCE# APP
--------------------- ---
15 YES
16 YES
17 YES
|
若在上步中的max sequence#在的的app状态为YES说明standby database 成功创建.
10. 以spfile启动并设为只读
SQL> create spfile from pfile;
SQL> shutdown immedaite
SQL> startup mount
SQL> alter database recover managed standby database disconnect from session;
SQL> alter database open read only;
|