高级复制也称为对称复制,分为多主体站点复制(Multiple Master Rplication)、物化视图站点复制(Materialized View Replication)和前两种的混合复制。高级复制主要是用在对称的、等同的数据库(仅限Oracle数据库之间)表之间的(单向或双向)的复制,以满足分布式应用的需求。多主体站点复制基于Oracle的触发器(Trigger)捕获DML和DDL操作,并将这些操作封装在Remote Procedure Calls(RPCs)里,并借助Deferred Transaction Queue来传递RPCs并在目标数据库利用Internal Trigger执行传递过来的RPCs从而实现了数据的同步,如果仅仅是高级复制环境中的只读物化视图环境,那么会直接利用解析物化视图日志的方式来实现数据的同步,而不用DTQ来传递RPCs,基于这一特点,高级复制对网络的稳定性和传输速度要求较高,配置也稍显麻烦。另外一点要特别注意,就是物化视图复制方式不支持DDL的变更同步。
流复制是Oracle 10g新推出的同步技术,基于logminer(数据库必须运行在Archive Log模式),通过Oracle Advanced Queue来实现数据的同步复制,可以实现表,用户,数据库级别的同步,而高级复制支持的对象仅为表,索引,同义词,触发器,视图,过程,函数,程序包,程序包体等。Stream相对高级复制而言配置简单,通过oracle gateway还可用于异构平台和异构数据库的数据同步,这是高级复制做不到的。
实验如下:
源端:
SYS@bys1>select name,db_unique_name from v$database;
NAME DB_UNIQUE_NAME
--------- ------------------------------
BYS1 bys1
SYS@bys1>show parameter global_name
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
global_names boolean TRUE
SYS@bys1>select * from global_name;
GLOBAL_NAME
----------------------------------------------------------------------------------------------------
BYS1
SYS@bys1>show parameter job_queue_processes
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
job_queue_processes integer 1000
#######
目标端:
SYS@bys2>select name,db_unique_name from v$database;
NAME DB_UNIQUE_NAME
------------------ ------------------------------------------------------------
BYS1 BYS2
SYS@bys2>show parameter global_name
NAME TYPE VALUE
------------------------------------ ---------------------- ------------------------------
global_names boolean TRUE
SYS@bys2>select * from global_name;
GLOBAL_NAME
----------------------------------------------------------------------------------------------------
BYS2
SYS@bys2>show parameter job_queue_processes
NAME TYPE VALUE
------------------------------------ ---------------------- ------------------------------
job_queue_processes integer 1000
###########################################################
复制用户--源和目标都做:
create user repadmin identified by repadmin;
默认的用户表空间是USER,默认临时表空间是TEMP的话,以下两句可以不执行。
alter user repadmin default tablespace users;
alter user repadmin temporary tablespace temp;
grant connect ,resource to repadmin;
execute dbms_repcat_admin.grant_admin_any_schema('REPADMIN');
grant comment any table to repadmin;
execute dbms_defer_sys.register_propagator('REPADMIN');
grant execute any procedure to REPADMIN;
源和目标端创建业务用户-源和目标都做:
create user gp identified by gp default tablespace users;
grant dba to gp;
###################################################################
SYS@bys1>conn gp/gp
Connected.
GP@bys1>create table test(x int primary key);
Table created.
GP@bys1>select * from tab;
TNAME TABTYPE CLUSTERID
------------------------------ ------- ----------
TEST TABLE
目标端:
SYS@bys2>conn gp/gp
Connected.
GP@bys2>create table test(x int primary key);
Table created.
###################################################################
注意:如果db_domain值为空,并且global names设置为true的情况下,那么这里link关键词后面的这个“bys2”必须写目标端global_name的值,而using关键词后面的“bys2”表示的是连接目标端的SERVICE NET NAME
源端:
GP@bys1>conn repadmin/repadmin
Connected.
REPADMIN@bys1>create database link bys2 connect to repadmin identified by repadmin using 'bys2';
Database link created.
REPADMIN@bys1>select * from global_name@bys2;
GLOBAL_NAME
-----------------------------------------------
BYS2
目标端:
GP@bys2>conn repadmin/repadmin
Connected.
REPADMIN@bys2>create database link bys1 connect to repadmin identified by repadmin using 'bys1';
Database link created.
REPADMIN@bys2>select * from global_name@bys1;
GLOBAL_NAME
---------------------------------------
BYS1
################################################################################
REPADMIN@bys1> execute dbms_repcat.create_master_repgroup('rep');
PL/SQL procedure successfully completed.
REPADMIN@bys1>select gname ,master ,status from dba_repgroup where gname = 'REP';
GNAME M STATUS
------------------------------ - ---------
REP Y QUIESCED
REPADMIN@bys1>execute dbms_repcat.create_master_repobject(sname=>'gp',oname=>'test',type=>'table',use_existing_object=>true ,gname=>'rep' ,copy_rows=>false);
PL/SQL procedure successfully completed.
REPADMIN@bys1>select gname ,master ,status from dba_repgroup where gname = 'REP';
GNAME M STATUS
------------------------------ - ---------
REP Y QUIESCED
REPADMIN@bys1>execute dbms_repcat.generate_replication_support('gp','test' ,'table');
PL/SQL procedure successfully completed.
REPADMIN@bys1>select gname ,master ,status from dba_repgroup where gname = 'REP';
GNAME M STATUS
------------------------------ - ---------
REP Y QUIESCED
col sname for a10
col oname for a10
col gname for a10
REPADMIN@bys1>select sname,oname,status,gname from dba_repobject; ---这一步要做一下检查,确认启动的复制对象是正常的。
SNAME ONAME STATUS GNAME
---------- ---------- ---------- ----------
GP TEST VALID REP
GP TEST$RP VALID REP
GP TEST$RP VALID REP
注:如添加源端复制支持这一步输入 错了表名或用户名,通过dba_repobject;可以查出,可以使用下面语句删除错误的配置,不然下一步添加复制节点会报错找不到相关对象。
execute DBMS_REPCAT.DROP_MASTER_REPOBJECT (sname=>'gp',oname=>'T',type=>'TABLE');
REPADMIN@bys1>execute dbms_repcat.add_master_database(gname=>'rep',master=>'bys2' ,use_existing_objects=>true ,copy_rows=>false ,propagation_mode=>'synchronous');
PL/SQL procedure successfully completed.
REPADMIN@bys1>col dblink for a10
REPADMIN@bys1> select gname ,dblink ,masterdef ,master from dba_repsites where gname='REP';
GNAME DBLINK M M
---------- ---------- - -
REP BYS1 Y Y
REP BYS2 N Y
REPADMIN@bys2>col gname for a10
REPADMIN@bys2>col dblink for a10
REPADMIN@bys2> select gname ,dblink ,masterdef ,master from dba_repsites where gname='REP';
GNAME DBLINK MA MA
---------- ---------- -- --
REP BYS1 Y Y
REP BYS2 N Y
REPADMIN@bys1>execute dbms_repcat.resume_master_activity('rep' ,true);
PL/SQL procedure successfully completed.
################################
源端:
13:51:48 GP@bys1>select * from test;
no rows selected
13:51:57 GP@bys1>insert into test values(2);
1 row created.
13:52:02 GP@bys1>commit;
Commit complete.
13:52:06 GP@bys1>select * from test;
X
----------
2
目标端:
13:51:19 GP@bys2>select * from test;
no rows selected
13:51:22 GP@bys2>select * from test;
no rows selected
13:51:52 GP@bys2>select * from test;
X
----------
2
13:52:13 GP@bys2>
源端:
13:52:08 GP@bys1>truncate table test;
Table truncated.
13:52:21 GP@bys1>select * from test;
no rows selected
目标端:
13:52:27 GP@bys2>select * from test;
X
----------
2
13:52:28 GP@bys2>select * from test;
X
----------
2