此文将在前两文基础上介绍如何配置一个一对多的数据分布式复制体系。

    方法与前面大同小异，再次仅突出重点。

    由于测试机器不够，我就在Target上的DB里建立两个User，分别代表两个Target环境。

    这里需要注意的是，两个Target可以使用同一个GG binary，也可以使用同一个MGR（同一个port），只是需要配置不同的REPLICAT而已。

1.Source 添加EXTRACT

GGSCI (xxx) 120> ADD EXTRACT HAOEXMUL, TRANLOG, BEGIN now
EXTRACT added.

GGSCI (xxx) 121> edit params HAOEXMUL

EXTRACT HAOEXMUL
SETENV (ORACLE_SID=MOT)
USERID GGS@MOT, PASSWORD GGS
EXTTRAIL ./dirdat/za
TABLE HAOZHU_USER.GGMULTI;
~
"dirprm/haoexmul.prm" 5 lines, 118 characters written

GGSCI (xxx) 124> ADD EXTTRAIL ./dirdat/za, EXTRACT HAOEXMUL
EXTTRAIL added.

GGSCI (xxx) 125> start extract HAOEXMUL

Sending START request to MANAGER ...
EXTRACT HAOEXMUL starting

GGSCI (xxx) 126> info HAOEXMUL

EXTRACT    HAOEXMUL  Last Started 2010-04-22 12:23   Status RUNNING
Checkpoint Lag       00:13:57 (updated 00:00:03 ago)
Log Read Checkpoint  Oracle Redo Logs
                     2010-04-22 12:09:54  Seqno 64, RBA 8503312

2.Source添加两个data pump EXTRACT

GGSCI (xxx) 110> ADD EXTRACT HAODP1, EXTTRAILSOURCE ./dirdat/za, BEGIN now
EXTRACT added.

GGSCI (xxx) 111> ADD EXTRACT HAODP2, EXTTRAILSOURCE ./dirdat/za, BEGIN now
EXTRACT added.

GGSCI (xxx) 112> ADD RMTTRAIL ./dirdat/za, EXTRACT HAODP1
RMTTRAIL added.

GGSCI (xxx) 113> ADD RMTTRAIL ./dirdat/zb, EXTRACT HAODP2
RMTTRAIL added.

GGSCI (xxx) 114> edit params HAODP1

EXTRACT HAODP1
SETENV (ORACLE_SID=MOT)
USERID GGS@MOT, PASSWORD GGS
RMTHOST xxx.xxx.xxx.xxx, MGRPORT 7809
RMTTRAIL ./dirdat/za
TABLE HAOZHU_USER.GGMULTI;
~

~
"dirprm/haodp1.prm" [New File] 6 lines, 158 characters written

GGSCI (xxx) 115> edit params HAODP2

EXTRACT HAODP2
SETENV (ORACLE_SID=MOT)
USERID GGS@MOT, PASSWORD GGS
RMTHOST xxx.xxx.xxx.xxx, MGRPORT 7809
RMTTRAIL ./dirdat/zb
TABLE HAOZHU_USER.GGMULTI;
~
"dirprm/haodp2.prm" [New File] 6 lines, 159 characters written

GGSCI (xxx) 128> start HAODP1

Sending START request to MANAGER ...
EXTRACT HAODP1 starting

GGSCI (xxx) 129> start HAODP2

Sending START request to MANAGER ...
EXTRACT HAODP2 starting

GGSCI (xxx) 130> info HAODP1

EXTRACT    HAODP1    Last Started 2010-04-22 12:29   Status RUNNING
Checkpoint Lag       00:00:00 (updated 00:00:03 ago)
Log Read Checkpoint  File ./dirdat/za000000
                     2010-04-22 12:04:39.000000  RBA 0

GGSCI (xxx) 131> info HAODP2

EXTRACT    HAODP2    Last Started 2010-04-22 12:29   Status RUNNING
Checkpoint Lag       00:00:00 (updated 00:00:02 ago)
Log Read Checkpoint  File ./dirdat/za000000
                     2010-04-22 12:04:44.000000  RBA 0

3.Target 添加两个Replicat

GGSCI (xxx) 32> ADD REPLICAT HAOREPA, EXTTRAIL ./dirdat/za, BEGIN now , nodbcheckpoint
REPLICAT added.

GGSCI (xxx) 33> ADD REPLICAT HAOREPB, EXTTRAIL ./dirdat/zb, BEGIN now , nodbcheckpoint
REPLICAT added.

GGSCI (xxx) 34> edit params HAOREPA

REPLICAT HAOREPA
ASSUMETARGETDEFS
SETENV (ORACLE_SID=DC1)
USERID GGS, PASSWORD GGS
map HAOZHU_USER.GGMULTI , target HAOZHU_USER.GGMULTI;
~

"dirprm/haorepa.prm" [New File] 5 lines, 137 characters written

GGSCI (xxx) 35> edit params HAOREPB

REPLICAT HAOREPB
ASSUMETARGETDEFS
SETENV (ORACLE_SID=DC1)
USERID GGS, PASSWORD GGS
map HAOZHU_USER.GGMULTI , target HAOZHU2_USER.GGMULTI;
~

"dirprm/haorepb.prm" [New File] 5 lines, 138 characters written

GGSCI (xxx) 36> start HAOREPA 

Sending START request to MANAGER ...
REPLICAT HAOREPA starting

GGSCI (xxx) 37> start HAOREPB

Sending START request to MANAGER ...
REPLICAT HAOREPB starting

GGSCI (xxx) 38> info HAOREPA

REPLICAT   HAOREPA   Last Started 2010-04-22 12:35   Status RUNNING
Checkpoint Lag       00:00:00 (updated 00:00:09 ago)
Log Read Checkpoint  File ./dirdat/za000000
                     2010-04-22 12:32:46.000000  RBA 0

GGSCI (xxx) 39> info HAOREPB

REPLICAT   HAOREPB   Last Started 2010-04-22 12:35   Status RUNNING
Checkpoint Lag       00:00:00 (updated 00:00:06 ago)
Log Read Checkpoint  File ./dirdat/zb000000
                     2010-04-22 12:32:50.000000  RBA 0

4.测试

在Source：

SQL> insert into ggmulti select * from dba_objects;

10045 rows created.

SQL> commit;

Commit complete.

在Target：

SQL> show user
USER is "HAOZHU2_USER"
SQL> select count(*) from ggmulti;

  COUNT(*)
----------
     10045

SQL> show user
USER is "HAOZHU_USER"
SQL> select count(*) from ggmulti;

  COUNT(*)
----------
     10045

（这里感谢Thomas Zhang的帮助让我找到一个配置中的失误:)）