GoldenGate12.2在DataGuard备库同步数据的配置

OGG 12.2.0.1在standby数据库上安装 

http://blog.itpub.net/10972173/viewspace-2638152/
###########################软件安装配置############################ 
1.设置ORACLE_HOME和ORACLE_SID(主库、备库、目标库) 
如果只有一个实例，只需要设置环境变量ORACLE_HOME和ORACLE_SID即可，如果有多个实例，需要在每个进程组(process group)的配置文件中设置： 

2.设置library的path,源端(主库、备库、目标库) 

export GG_HOME=/home/oracle/ogg 
export PATH=$PATH:$GG_HOME 
export LIBPATH=$GG_HOME:$ORACLE_HOME/lib 

3.执行runinstaller(主库、备库、目标库) 
./runinstaller 

4.修改日志模式(主库) 

SELECT supplemental_log_data_min, force_logging FROM v$database; 
ALTER DATABASE ADD SUPPLEMENTAL LOG DATA; 
ALTER DATABASE FORCE LOGGING; 
SELECT supplemental_log_data_min, force_logging FROM v$database; 

5.在数据库中启用ogg(主库、目标库) 

alter system set ENABLE_GOLDENGATE_REPLICATION=true scope=both; 

6.设置Flashback Query(主库和备库) 

UNDO_MANAGEMENT=AUTO 
UNDO_RETENTION=86400 #需要计算得出 

8.建立GoldenGate Credentials(主库、目标库) 

create tablespace ogg_tbs datafile '/home/db/oracle/oradata/bidb/ogg_tbs01.dbf' size 1m autoextend on next 1m maxsize unlimited; 
create user ogg identified by ogg123 default tablespace ogg_tbs quota unlimited on ogg_tbs; 

9.执行setup脚本(主库，目标库) 
@marker_setup.sql 
@ddl_setup.sql 
@role_setup.sql 
@ddl_enable.sql 
@ddl_pin ogg 

10.安全相关认证(主库、备库、目标库) 
 $ ./ggsci 
注意下面两个语句末尾不能有分号，否则报ERROR: Invalid command. 
GGSCI> add credentialstore 
GGSCI> alter credentialstore add user ogg,password ogg123 
GGSCI> dblogin useridalias ogg 

########################主库上添加trandata############################### 
11.在DG主库上添加trandata(主库) 
 $ ./ggsci 
GGSCI> dblogin useridalias ogg 
GGSCI> add trandata scott.T 
GGSCI> info trandata scott.* 

#######################源端配置################################ 
12 配置manager进程(备库) 
 $ ./ggsci 
GGSCI> edit params mgr 
内容如下： 
PORT 7809 
DYNAMICPORTLIST 7801-7809 
--AUTORESTART ER *,RETRIES 5,WAITMINUTES 7 
PURGEOLDEXTRACTS ./dirdat/*,usecheckpoints, minkeepdays 7 
LAGREPORTHOURS 1 
LAGINFOMINUTES 30 
LAGCRITICALMINUTES 45 


13.配置抽取进程： 
GGSCI> add extract extncf, tranlog, begin now 
GGSCI> add exttrail ./dirdat/et extract extncf, megabytes 100 
GGSCI> edit params extncf 
内容如下： 
EXTRACT extncf 
--SETENV (ORACLE_HOME="/home/db/oracle/product/11.2.0") 
--setenv (NLS_LANG="AMERICAN_AMERICA.AL32UTF8") 
--setenv (ORACLE_SID="stby") 
useridalias ogg 
GETTRUNCATES 
REPORTCOUNT EVERY 1 MINUTES, RATE 
DISCARDFILE ./dirrpt/extncf.dsc,APPEND,MEGABYTES 1000 

WARNLONGTRANS 2h,CHECKINTERVAL 10m 
EXTTRAIL ./dirdat/et 

--TRANLOGOPTIONS  CONVERTUCS2CLOBS 
TRANLOGOPTIONS EXCLUDEUSER ogg 
TRANLOGOPTIONS MINEFROMACTIVEDG 
DBOPTIONS ALLOWUNUSEDCOLUMN 
--DYNAMICRESOLUTION 

FETCHOPTIONS FETCHPKUPDATECOLS 

--table 
table SCOTT.T; 

14.配置datapump进程 
GGSCI> add extract dpencf, exttrailsource ./dirdat/et 
GGSCI> add rmttrail ./dirdat/rt, extract dpencf, megabytes 100 
GGSCI> edit params dpencf 
内容如下： 
EXTRACT dpencf 
RMTHOST 192.168.137.7, MGRPORT 7800, compress 
PASSTHRU 
RMTTRAIL ./dirdat/rt 
DYNAMICRESOLUTION 

--table 
table SCOTT.T; 


15.启动进程 
 $ ./ggsci 
GGSCI> start mgr 
GGSCI> start extncf 

#######################目标配置################################ 
16.目标端GLOBALS配置： 
GGSCI> edit params ./GLOBALS 
内容如下： 
CHECKPOINTTABLE ogg.oggchkpt 
GGSCI> exit 
Shell> ggsci 
（重新登陆以激活GLOBALS参数） 
GGSCI> dblogin useridalias ogg 
GGSCI> add checkpointtable 

17 配置目标端manager进程： 
GGSCI> edit params mgr 
内容如下： 
port 7800 
DYNAMICPORTLIST 7801-7809 
--AUTORESTART ER *,RETRIES 5,WAITMINUTES 7 
PURGEOLDEXTRACTS ./dirdat/*,usecheckpoints, minkeepdays 7 
LAGREPORTHOURS 1 
LAGINFOMINUTES 30 
LAGCRITICALMINUTES 45 

18 配置目标端复制进程： 
GGSCI> add replicat repncf, exttrail ./dirdat/rt 
GGSCI> edit params repncf 
内容如下： 
REPLICAT repncf 
SETENV (ORACLE_HOME="/home/db/oracle/product/11.2.0") 
setenv (NLS_LANG="AMERICAN_AMERICA.AL32UTF8") 
setenv (ORACLE_SID="bidb") 
useridalias ogg 
--SQLEXEC "ALTER SESSION SET CONSTRAINTS=DEFERRED" 
REPORT AT 01:59 
REPORTCOUNT EVERY 30 MINUTES, RATE 
REPERROR DEFAULT, ABEND 
--numfiles 5000 

--HANDLECOLLISIONS 
assumetargetdefs 
DISCARDFILE ./dirrpt/repncf.dsc, APPEND, MEGABYTES 1000 
GETTRUNCATES 
ALLOWNOOPUPDATES 

--table 
map SCOTT.T, target SCOTT.T; 

#######################数据初始化################################ 
19.源端导出数据(主库操作) 
SQL> select dbms_flashback.get_system_change_number from dual; 

GET_SYSTEM_CHANGE_NUMBER 
------------------------ 
        347494 

expdp \'/ as sysdba\' directory=dump_dir dumpfile=scott.dmp logfile=scott.log tables=scott.t  flashback_scn=347494 
scp scott.dmp m3:/home/oracle/dump_dir 

20.目标端导入 
cd /home/oracle/dump_dir 
impdp \'/ as sysdba\' directory=dump_dir dumpfile=scott.dmp logfile=scott.log 


#######################启动进程################################ 
20.启动进程 
源端： 
GGSCI> start extncf 
目标端： 
GGSCI> alter replicat repncf extseqno 0, extrba 0 
此处的extseqno后面的0代表trail文件的序号，需要到ogg根目录下的dirdat中检查，如果初始配置的话，就是0 
GGSCI> start repncf aftercsn 347494 

 

APPLIES TO:

Oracle GoldenGate - Version 11.2.1.0.37 and later
Information in this document applies to any platform.

GOAL

How To Configure Extract On Active Data Guard If Primary Is 2 Or More Node RAC And DG Is Non RAC.
 

SOLUTION

To Configure capture on Active Data Guard if primary is 2 node RAC and DG is non RAC kindly refer below :

From the Active Data Guard(ADG) perspective, when you configure Active Data Guard standby in a single instance database, you should create 2 threads of standby redo logs if the primary database is 2-node RAC.
It will capture both DDL and DML.But please check the limitation on below documentation :

http://docs.oracle.com/goldengate/1212/gg-winux/GIORA/classic_capture.htm#A1740233


Topic " 11.6.1 Limitations and Requirements for Using ADG Mode "


Example :

Primary (2-node RAC)
(Online Redo Log)
- thread 1 (group 1, group 2, group 3)
- thread 2 (group 4, group 5, group 6)

Standby (Single Instance)
(Online Redo Log)
- thread 1 (group 1, group 2, group 3)
(Standby Redo Log)
- thread 1 (group 1, group 2, group 3)
- thread 2 (group 4, group 5, group 6)