Oracle 流是一种数据共享的通用机制，可以被用于许多处理的基础，包括消息、复制和数据仓库的 ETL 过程。它是高级队列、LogMinor、作业调度等已存在技术的扩展。这里做一个简单的复制示例来说明其用法。试验环境：

windows server 2003 sp1；10g 10.1.0.2
源库：SID：db1 IP：10.1.8.201
目的库：SID：db2 IP：10.1.9.49

源、目的库的 tnsnames.ora 配置
DB1 =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = 10.1.8.201)(PORT = 1521))
(CONNECT_DATA =
(SERVICE_NAME = db1)
(SERVER = DEDICATED)
)
)

DB2 =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = 10.1.9.49)(PORT = 1521))
)
(CONNECT_DATA =
(SERVICE_NAME = db2)
(SERVER = DEDICATED)
)
)

1. 基本体系结构

流处理分为捕捉、传输、应用三个主要进程。

捕捉进程是一个可选的后台进程。它从重做日志中捕获 DDL 和 DML 的变化，并且把它们封装成逻辑改变记录（LCRs）。缺省的捕捉机制可以立即获得用户定义的事件。
传输进程把 LCRs 存储到 SYS.AnyData 数据类型的队列中。LCRs 在不同数据库中的源和目的传输区域之间传播。传播使用工作队列来调度。
应用进程是一个可选的后台进程。它调出 LCRs 直接应用，或者为用户自定义消息作为参数传递给用户自定义包。

捕获和应用进程可以被用于表、模式、数据库等级别，各级别的规则决定了其不同的行为。

2. 实例设置

CONN sys/password@DB1 AS SYSDBA
create pfile from spfile;
-- 编辑 initsid.ora 文件，增加以下参数：
-- AQ_TM_PROCESSES=1;
-- GLOBAL_NAMES=TRUE;

SHUTDOWN IMMEDIATE;
create spfile from pfile;
STARTUP mount;
alter database archivelog; -- 改为归档模式
alter database open;

3. 流管理员设置

CONN sys/password@DB1 AS SYSDBA

CREATE USER strmadmin IDENTIFIED BY strmadmin DEFAULT TABLESPACE users QUOTA UNLIMITED ON users;

GRANT DBA, CONNECT, RESOURCE, SELECT_CATALOG_ROLE TO strmadmin;

GRANT EXECUTE ON DBMS_AQADM TO strmadmin;
GRANT EXECUTE ON DBMS_CAPTURE_ADM TO strmadmin;
GRANT EXECUTE ON DBMS_PROPAGATION_ADM TO strmadmin;
GRANT EXECUTE ON DBMS_STREAMS_ADM TO strmadmin;
GRANT EXECUTE ON DBMS_APPLY_ADM TO strmadmin;
GRANT EXECUTE ON DBMS_FLASHBACK TO strmadmin;

GRANT ALL ON scott.dept TO strmadmin;

BEGIN
DBMS_RULE_ADM.GRANT_SYSTEM_PRIVILEGE(
privilege => DBMS_RULE_ADM.CREATE_RULE_SET_OBJ,
grantee => 'strmadmin',
grant_option => FALSE);
END;
/

BEGIN
DBMS_RULE_ADM.GRANT_SYSTEM_PRIVILEGE(
privilege => DBMS_RULE_ADM.CREATE_RULE_OBJ,
grantee => 'strmadmin',
grant_option => FALSE);
END;
/

CONNECT strmadmin/strmadmin@DB1
EXEC DBMS_STREAMS_ADM.SET_UP_QUEUE(); -- 建立流队列

CREATE DATABASE LINK db2 CONNECT TO strmadmin IDENTIFIED BY strmadmin USING 'DB2'; -- 建立数据库连接

在目的库（DB2）重复上面2、3步骤。

4. LogMinor 表空间设置

CONN sys/password@DB1 AS SYSDBA

CREATE TABLESPACE logmnr_ts DATAFILE 'E:ORACLEPRODUCT10.1.0ORADATADB1logmnr01.dbf' SIZE 25 M REUSE AUTOEXTEND ON MAXSIZE UNLIMITED;

EXECUTE DBMS_LOGMNR_D.SET_TABLESPACE('logmnr_ts');

5. SUPPLEMENTAL 日志

CONN sys/password@DB1 AS SYSDBA
ALTER TABLE scott.dept ADD SUPPLEMENTAL LOG GROUP log_group_dept_pk (deptno) ALWAYS;

6. 配置传播进程

CONNECT strmadmin/strmadmin@DB1
BEGIN
DBMS_STREAMS_ADM.ADD_TABLE_PROPAGATION_RULES(
table_name => 'scott.dept',
streams_name => 'db1_to_db2',
source_queue_name => 'strmadmin.streams_queue',
destination_queue_name => 'strmadmin.streams_queue@db2',
include_dml => true,
include_ddl => true,
source_database => 'db1');
END;
/

-- 检查传播进程作业
SELECT job,TO_CHAR(last_date, 'DD-Mon-YYYY HH24:MI:SS') last_date,TO_CHAR(next_date, 'DD-Mon-YYYY HH24:MI:SS') next_date,what FROM dba_jobs;

7. 配置捕捉进程

CONNECT strmadmin/strmadmin@DB1
BEGIN
DBMS_STREAMS_ADM.ADD_TABLE_RULES(
table_name => 'scott.dept',
streams_type => 'capture',
streams_name => 'capture_simp',
queue_name => 'strmadmin.streams_queue',
include_dml => true,
include_ddl => true);
END;
/

8. 配置初始 SCN
在应用进程工作前必须在目的表中配置源表的 SCN。如果目的表已经存在，可用 exp/imp 元数据完成。

exp userid=scott/tiger@db1 FILE=dept_instant.dmp TABLES=dept OBJECT_CONSISTENT=y ROWS=n
imp userid=scott/tiger@db2 FILE=dept_instant.dmp IGNORE=y COMMIT=y LOG=import.log STREAMS_INSTANTIATION=y

因为在迁移元数据时 SUPPLEMENTAL 日志也被迁移。因为捕捉进程还没有启动，所以可以删除这些日志：

CONN sys/password@DB2 AS SYSDBA

ALTER TABLE scott.dept DROP SUPPLEMENTAL LOG GROUP log_group_dept_pk;

用 DBMS_APPLY_ADM 包设置 SCN

CONNECT strmadmin/strmadmin@db1

DECLARE

v_scn NUMBER;

BEGIN

v_scn := DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER();

DBMS_APPLY_ADM.SET_TABLE_INSTANTIATION_SCN@DB2(

source_object_name => 'scott.dept',

source_database_name => 'db1',

instantiation_scn => v_scn);

END;

/

9. 配置应用进程

CONNECT strmadmin/strmadmin@DB2
BEGIN
DBMS_STREAMS_ADM.ADD_TABLE_RULES(
table_name => 'scott.dept',
streams_type => 'apply',
streams_name => 'apply_simp',
queue_name => 'strmadmin.streams_queue',
include_dml => true,
include_ddl => true,
source_database => 'db1');
END;
/

10. 启动应用进程

CONNECT strmadmin/strmadmin@DB2
BEGIN
DBMS_APPLY_ADM.SET_PARAMETER(
apply_name => 'apply_simp',
parameter => 'disable_on_error',
value => 'n');

DBMS_APPLY_ADM.START_APPLY(
apply_name => 'apply_simp');
END;
/

11. 启动捕捉进程

CONNECT strmadmin/strmadmin@DB1
BEGIN
DBMS_CAPTURE_ADM.START_CAPTURE(
capture_name => 'capture_simp');
END;
/

12. 测试

-- 测试 DML

CONNECT scott/tiger@db1

INSERT INTO dept (deptno, dname, loc) VALUES (99, 'Test Dept', 'UK');

COMMIT;

SELECT * FROM dept;

DEPTNO DNAME LOC

---------- -------------- -------------

10 ACCOUNTING NEW YORK

20 RESEARCH DALLAS

30 SALES CHICAGO

40 OPERATIONS BOSTON

99 Test Dept UK

5 rows selected.

CONNECT scott/tiger@db2

SELECT * FROM dept;

DEPTNO DNAME LOC

---------- -------------- -------------

10 ACCOUNTING NEW YORK

20 RESEARCH DALLAS

30 SALES CHICAGO

40 OPERATIONS BOSTON

99 Test Dept UK

5 rows selected.

-- 测试 DDL

CONNECT scott/tiger@db1

ALTER TABLE dept ADD (new_col NUMBER(10))

/

DESC dept

Name Null? Type

---------------------------- -------- --------------

DEPTNO NOT NULL NUMBER(2)

DNAME VARCHAR2(14)

LOC VARCHAR2(13)

NEW_COL NUMBER(10)

CONNECT scott/tiger@db2

DESC dept

Name Null? Type

---------------------------- -------- --------------

DEPTNO NOT NULL NUMBER(2)

DNAME VARCHAR2(14)

LOC VARCHAR2(13)

NEW_COL NUMBER(10)

-- 可以用下列语句查看流的内容

CONNECT strmadmin/strmadmin@DB1

SELECT s.user_data.getTypeName() FROM streams_queue_table s;

SET SERVEROUTPUT ON

DECLARE

v_anydata SYS.ANYDATA;

v_lcr SYS.LCR$_ROW_RECORD;

v_row_list SYS.LCR$_ROW_LIST;

v_result PLS_INTEGER;

BEGIN

SELECT user_data

INTO v_anydata

FROM strmadmin.streams_queue_table

WHERE rownum < 2;

v_result := ANYDATA.GetObject(

self => v_anydata,

obj => v_lcr);

DBMS_OUTPUT.PUT_LINE('Command Type : ' || v_lcr.Get_Command_Type);

DBMS_OUTPUT.PUT_LINE('Object Owner : ' || v_lcr.Get_Object_Owner);

DBMS_OUTPUT.PUT_LINE('Object Name : ' || v_lcr.Get_Object_Name);

DBMS_OUTPUT.PUT_LINE('Source Database Name : ' || v_lcr.Get_Source_Database_Name);

END;

/

13. 清理

-- 可以使用下面的语句识别并删除所有规则
conn / as sysdba

BEGIN
FOR cur_rec IN (SELECT rule_owner, rule_name FROM dba_rules) LOOP
DBMS_RULE_ADM.DROP_RULE(
rule_name => cur_rec.rule_owner || '.' || cur_rec.rule_name,
force => TRUE);
END LOOP;
END;
/

-- 识别、停止并删除所有捕捉进程和应用进程
conn / as sysdba

BEGIN
FOR cur_rec IN (SELECT capture_name FROM dba_capture) LOOP
DBMS_CAPTURE_ADM.STOP_CAPTURE(
capture_name => cur_rec.capture_name);
DBMS_CAPTURE_ADM.DROP_CAPTURE(
capture_name => cur_rec.capture_name);
END LOOP;

FOR cur_rec IN (SELECT apply_name FROM dba_apply) LOOP
DBMS_APPLY_ADM.STOP_APPLY(
apply_name => cur_rec.apply_name);
DBMS_APPLY_ADM.DROP_APPLY(
apply_name => cur_rec.apply_name);
END LOOP;
END;
/

-- 使用下面的语句彻底删除相关对象的流信息
BEGIN
DBMS_STREAMS_ADM.PURGE_SOURCE_CATALOG(
source_database => 'db1',
source_object_name => 'scott.dept',
source_object_type => 'TABLE');
END;
/

ITPub博客

Oracle 10g stream 一对多复制

最新文章