oracle dataguard 实验(单数据库环境+物理standby)
本次oracle dataguard
环境：
操作系统：windows 2003 server
数据库：oracle 10g 10.2.0.1
ORACLE_HOME：D:/oracle/product/10.2.0/db_1
archive_dest：D:/archivelog
rman_dest：d:/rman_backup
机器：1台
主库名称：learn
备库名称：learndg
实验步骤：修改好tnsnames、listener、pfile文件，通过rman的duplicate还原备库
一、主库上操作：
alter database force logging;
alter database archivelog;
添加standby日志组，比本身的redo日志组多出一组（有一个推荐的公式可以做参考：(每线程的日志组数+1)*最大线程数 线程数可以理解为rac结构中的rac节点数）
alter database add standby logfile group 4 ('D:/oracle/product/10.2.0/oradata/learn/STANDBYRDO01.LOG') szie 50M;
create pfile='d:/oracle/init_learn.ora' from spfile;
复制D:/oracle/product/10.2.0/db_1/database目录下的PWDlearn.ora为PWDlearndg.ora
修改create出来的pfile修改后如下,红字部分为修改的地方（里面没有的则添加）
learn.__db_cache_size=276824064
learn.__java_pool_size=4194304
learn.__large_pool_size=4194304
learn.__shared_pool_size=109051904
learn.__streams_pool_size=0
*.audit_file_dest='d:/oracle/product/10.2.0/admin/learn/adump'
*.background_dump_dest='d:/oracle/product/10.2.0/admin/learn/bdump'
*.compatible='10.2.0.1.0'
*.control_files='d:/oracle/product/10.2.0/oradata/learn//control01.ctl','d:/oracle/product/10.2.0/oradata/learn//control02.ctl','d:/oracle/product/10.2.0/oradata/learn//control03.ctl'
*.core_dump_dest='d:/oracle/product/10.2.0/admin/learn/cdump'
*.db_block_size=8192
*.db_domain=''
*.db_file_multiblock_read_count=16
*.DB_FILE_NAME_CONVERT='D:/oracle/product/10.2.0/oradata/learndg','D:/oracle/product/10.2.0/oradata/learn'
*.db_name='learn'
*.db_recovery_file_dest='d:/oracle/product/10.2.0/flash_recovery_area'
*.db_recovery_file_dest_size=536870912
*.DB_UNIQUE_NAME='learn'
*.dispatchers='(PROTOCOL=TCP) (SERVICE=learnXDB)'
*.FAL_CLIENT='learn'
*.FAL_SERVER='learndg'
*.job_queue_processes=10
*.LOG_ARCHIVE_CONFIG='DG_CONFIG=(learn,learndg)'
*.log_archive_dest_1='LOCATION=D:/archive_log VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=learn'
*.LOG_ARCHIVE_DEST_2='SERVICE=learndg LGWR ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=learndg'
*.LOG_FILE_NAME_CONVERT='D:/oracle/product/10.2.0/oradata/learndg','D:/oracle/product/10.2.0/oradata/learn'
*.nls_language='SIMPLIFIED CHINESE'
*.nls_territory='CHINA'
*.open_cursors=300
*.pga_aggregate_target=134217728
*.processes=50
*.REMOTE_LOGIN_PASSWORDFILE='EXCLUSIVE'
*.sessions=60
*.sga_max_size=402653184
*.sga_target=402653184
*.STANDBY_FILE_MANAGEMENT='AUTO'
*.undo_management='AUTO'
*.undo_tablespace='UNDOTBS1'
*.user_dump_dest='d:/oracle/product/10.2.0/admin/learn/udump'
修改完成后
shutdown immediate
create spfile from pfile='d:/oracle/init_learn.ora';
startup
二、备库操作
复制刚才的'd:/oracle/init_learn.ora'为'd:/oracle/init_learndg.ora'
修改init_learndg.ora文件如下，红字部分为修改的地方
learn.__db_cache_size=276824064
learn.__java_pool_size=4194304
learn.__large_pool_size=4194304
learn.__shared_pool_size=109051904
learn.__streams_pool_size=0
*.audit_file_dest='d:/oracle/product/10.2.0/admin/learndg/adump'
*.background_dump_dest='d:/oracle/product/10.2.0/admin/learndg/bdump'
*.compatible='10.2.0.1.0'
*.control_files='d:/oracle/product/10.2.0/oradata/learndg//standbyctl01.ctl','d:/oracle/product/10.2.0/oradata/learndg//standbyctl02.ctl','d:/oracle/product/10.2.0/oradata/learndg//standbyctl03.ctl'
*.core_dump_dest='d:/oracle/product/10.2.0/admin/learndg/cdump'
*.db_block_size=8192
*.db_domain=''
*.db_file_multiblock_read_count=16
*.DB_FILE_NAME_CONVERT='D:/oracle/product/10.2.0/oradata/learn','D:/oracle/product/10.2.0/oradata/learndg'
*.LOG_FILE_NAME_CONVERT='D:/oracle/product/10.2.0/oradata/learn','D:/oracle/product/10.2.0/oradata/learndg'
*.db_name='learn'
*.db_recovery_file_dest='d:/oracle/product/10.2.0/flash_recovery_area'
*.db_recovery_file_dest_size=536870912
*.DB_UNIQUE_NAME='learndg'
*.dispatchers='(PROTOCOL=TCP) (SERVICE=learnXDB)'
*.FAL_CLIENT='learndg'
*.FAL_SERVER='learn'
*.job_queue_processes=10
*.LOG_ARCHIVE_CONFIG='DG_CONFIG=(learn,learndg)'
*.log_archive_dest_1='LOCATION=D:/archive_log/learndg/ VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=learndg'
*.LOG_ARCHIVE_DEST_2='SERVICE=learn LGWR ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=learn'
*.nls_language='SIMPLIFIED CHINESE'
*.nls_territory='CHINA'
*.open_cursors=300
*.pga_aggregate_target=134217728
*.processes=50
*.REMOTE_LOGIN_PASSWORDFILE='EXCLUSIVE'
*.sessions=60
*.sga_max_size=402653184
*.sga_target=402653184
*.STANDBY_FILE_MANAGEMENT='AUTO'
*.undo_management='AUTO'
*.undo_tablespace='UNDOTBS1'
*.user_dump_dest='d:/oracle/product/10.2.0/admin/learndg/udump'
在D:/oracle/product/10.2.0/admin目录下建立learndg文件夹，在learndg文件夹中建立adump bdump cdump dpump pfile udump文件夹
CMD下执行
oradim -new -sid LEARNDG
set ORACLE_SID=learndg
sqlplus /@learndg as sysdba
create spfile from pfile='d:/oracle/init_learn.ora';
startup nomount
quit
三、修改tnsnames、listener文件
tnsnames文件增加服务名learn和learndg，修改后如下
# tnsnames.ora Network Configuration File: d:/oracle/product/10.2.0/db_1/network/admin/tnsnames.ora
# Generated by Oracle configuration tools.
LEARNDG =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.53.26)(PORT = 1521))
)
(CONNECT_DATA =
(SERVICE_NAME = learndg)
)
)
LEARN =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.53.26)(PORT = 1521))
)
(CONNECT_DATA =
(SERVICE_NAME = learn)
)
)
EXTPROC_CONNECTION_DATA =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC0))
)
(CONNECT_DATA =
(SID = PLSExtProc)
(PRESENTATION = RO)
)
)
修改listener.ora文件，修改好的listener.ora如下
# listener.ora Network Configuration File: d:/oracle/product/10.2.0/db_1/network/admin/listener.ora
# Generated by Oracle configuration tools.
SID_LIST_LISTENER =
(SID_LIST =
(SID_DESC =
(SID_NAME = PLSExtProc)
(ORACLE_HOME = d:/oracle/product/10.2.0/db_1)
(PROGRAM = extproc)
)
(SID_DESC =
(GLOBAL_NAME = LEARNDG)
(ORACLE_HOME = d:/oracle/product/10.2.0/db_1)
(SID_NAME = learndg)
)
)
LISTENER =
(DESCRIPTION_LIST =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = IT0001)(PORT = 1521))
(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC0))
)
)
修改完成后
CMD命令下
lsnrctl stop
lsnrctl start
四、通过rman备份数据库，并使用duplicate同步出备库
rman nocatalog target /
run{
allocate channel d1 type disk format 'd:/rman_backup/%U';
allocate channel d2 type disk format 'd:/rman_backup/%U';
backup database skip inaccessible filesperset 10 plus archivelog filesperset 20 include current controlfile for standby delete all input;
release channel d1;
release channel d2;
}
备份完成后
connect auxiliary /@learndg
run{
duplicate target database for standby dorecover
db_file_name_convert=('D:/oracle/product/10.2.0/oradata/learn','D:/oracle/product/10.2.0/oradata/learndg');
}
五、rman恢复完成后
在备库上执行 alter database recover managed standby database disconnect from session;
在主库上执行 alter system switch logfile;
在主库和备库上执行
select max(sequence#) from v$archived_log;来验证是否同步成功
总结：整个实验过程历时4天，中途遇到如下问题，导致花费很多时间
1.监听问题、连接备库时提示
ORA-12154: TNS: 无法解析指定的连接标识符，原因是learndg由于无法启动注册监听，所以需要配置静态监听。
2.rman的duplicate问题，最初不知道duplicate for standby命令，直接通过duplicate命令同步备库，结果导致主库生成的standby controlfile和数据库文件不一致（数据库ID不同）
3.rman duplicate时在执行内存脚本这一步卡住，然后提示
RMAN-06136: 来自辅助数据库的 ORACLE 错误: ORA-01013: 用户请求取消当前的操作，原因是做duplicate时备库上还有用户登入，退出登入后错误解决
4.rman duplicate时提示
RMAN-03002: failure of Duplicate Db command at 12/14/2007 17:07:13
RMAN-03015: error occurred in stored script Memory Script
RMAN-06026: some targets not found - aborting restore
RMAN-06024: no backup or copy of the controlfile found to restore
原因是我之前的脚本是
run{
allocate channel d1 type disk format 'd:/rman_backup/%U';
allocate channel d2 type disk format 'd:/rman_backup/%U';
backup database skip inaccessible filesperset 10 plus archivelog filesperset 20 delete all input;
backup current controlfile for standby;
release channel d1;
release channel d2;
}
在自己的电脑上没问题，公司电脑上出现这个错
将脚本修改为如下后，问题解决
run{
allocate channel d1 type disk format 'd:/rman_backup/%U';
allocate channel d2 type disk format 'd:/rman_backup/%U';
backup database skip inaccessible filesperset 10 plus archivelog filesperset 20 include current controlfile for standby delete all input;
release channel d1;
release channel d2;
}