oracle stream参数,oracle stream实验(双向复制)

环境：

OS:Redhat Linux

Oracle 11.2.0.1

source:TEST

destnation:STANDBY

==准备==

1.归档日志配置：

SQL> archive log list;

Database log mode Archive Mode

Automatic archival Enabled

Archive destination /home/oracle/archivelog

Database log mode Archive Mode

Automatic archival Enabled

Archive destination /home/oracle/archivelog

---以下没有专门说明的都是sys用户---

2.在源stream1上配置Supplemental loging：

在源上：

SQL> alter database add supplemental log data;

Database altered.

3.两台机器上修改global_name:

stream1:

SQL> alter system set global_names=true scope=both;

SQL> alter database rename global_name to TEST.com;

stream2:

SQL> alter system set global_names=true scope=both;

SQL> alter database rename global_name to STANDBY.com;

4.修改初始化参数：

参考官方文档确定需要来检查和设定参数

----------ref 01----------------------------------------

在主/从数据库设置初始化参数(在主从数据库都执行以下操作)

1) sqlplus / as sysdba ----以dba身份登录

2)alter system set aq_tm_processes=2 scope=both;   ---启用对队列消息的时间监视

alter system set global_names=true scope=both;   ---设置全局名称为true

--alter system set job_queue_processes=10 scope=both; --设置定时作业进程最大为10个,默认是10

alter system set undo_retention=3600 scope=both;   --设置回滚段时间,默认是900

alter system set streams_pool_size=25M scope=spfile; --sga设置为自动调整情况下不需设置该参

--------------ref 02-----------------------------------------

源数据库和目的数据库均需要设置的参数：

alter system set global_names=true scope = both;

默认为false,  Database Link 使用的是数据库的global_name。

alter system set aq_tm_processes=2 scope=both;

以下参数都是10G的默认值， 检查下就可以了. 无需设置

如：Show parameter job_queue_processes

alter system set job_queue_processes = 10 scope=both;

alter system set sga_target = 300m scope=spfile;

alter system set open_links=4 scope=spfile;

alter system set statistics_level='TYPICAL' scope=both;

10g 默认为Typical， 性能统计模式

alter system set logmnr_max_persistent_sessions=1 scope=spfile;

10g 默认为1，持久的日志挖掘会话数。

alter system set "_job_queue_interval"=1 scope=spfile;

alter system set aq_tm_processes=1;

alter system set streams_pool_size=200m scope=both;

注意streams_pool_size一定要够大，因为如果启用了SGA_TARGET，ORACLE可能分配很少内存给stream导致大量信息被spill到磁盘导致查询DBA_APPLY,DBA_CAPTURE,DBA_PROPGATION全部状态ENABLED但就是没有数据被同步。

同时设置_job_queue_interval也是为了提高队列检查时间，防止apply出问题。

注意：与复制有关的2个参数：

如果等了足够长的时间发现数据没有复制过来，仔细检查了capture/propagation/apply各进程的状态都是正常的， 并检查参数.

alter system set "_job_queue_interval"=1 scope=spfile;

并且将aq_tm_processes参数改为1(我原来这是为10)

alter system set aq_tm_processes=1;

改完后重启，发现数据就可以去了。这个隐含参数只是控制对job队列的检查频率，默认5秒。

5.创建stream管理用户并表空间，配置权限，source和dest：

5.1 为strmadmin用户创建独立表空间

TEST:

SQL> CREATE TABLESPACE streams_tbs DATAFILE '/oracle/oradata/TEST/streams_tbs.dbf' SIZE 25M REUSE AUTOEXTEND ON MAXSIZE UNLIMITED;

Tablespace created.

STANDBY:

SQL> CREATE TABLESPACE streams_tbs DATAFILE '/oracle/oradata/STANDBY/streams_tbs.dbf' SIZE 25M REUSE AUTOEXTEND ON MAXSIZE UNLIMITED;

Tablespace created.

5.2 source和dest创建相同用户和权限：

SQL> CREATE USER strmadmin IDENTIFIED BY strmadminpw

DEFAULT TABLESPACE streams_tbs

QUOTA UNLIMITED ON streams_tbs;

User created.

SQL> GRANT DBA TO strmadmin;

Grant succeeded.

SQL>

BEGIN

DBMS_STREAMS_AUTH.GRANT_ADMIN_PRIVILEGE(

grantee => 'strmadmin',

grant_privileges => true);

END;

/

PL/SQL procedure successfully completed.

6.配置listener.ora和tnsnames.ora保证：

TEST 上：sqlplus strmadmin/strmadminpw@STANDBY

STANDBY 上：sqlplus strmadmin/strmadminpw@TEST

能够连通

---如果没有说明以下都是strmadmin/strmadminpw 用户---

7.创建DB link:

单向复制之需要一个database link即source 到dest:TEST-->STANDBY

TEST上：

sqlplus strmadmin/strmadminpw

SQL> CREATE DATABASE LINK STANDBY.com CONNECT TO strmadmin IDENTIFIED BY strmadminpw USING 'STANDBY';

Database link created.

测试：

SQL> select * from global_name@STANDBY.com;

GLOBAL_NAME

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

STANDBY.com

==开始stream的配置===

8. source和dest 创建队列：

TEST,STANDBY都要创建

SQL> EXEC DBMS_STREAMS_ADM.SET_UP_QUEUE();

PL/SQL procedure successfully completed.

上面命令会创建一个队列缺省名：streams_queue，队列表缺省是：STREAMS_QUEUE_TABLE

队列存储的object类型是anaydata

可以用查询dba_queues,dba_queue_tables来检查：

SQL> select owner,queue_table,name from dba_queues where owner='STRMADMIN';

OWNER QUEUE_TABLE NAME

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

STRMADMIN STREAMS_QUEUE_TABLE STREAMS_QUEUE

STRMADMIN STREAMS_QUEUE_TABLE AQ$_STREAMS_QUEUE_TABLE_E

SQL> select owner,queue_table,object_type from dba_queue_tables where owner='STRMADMIN';

OWNER QUEUE_TABLE OBJECT_TYPE

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

STRMADMIN STREAMS_QUEUE_TABLE SYS.ANYDATA

9. 在source: TEST上创建Stream propagation：

SQL>

BEGIN

DBMS_STREAMS_ADM.ADD_SCHEMA_PROPAGATION_RULES(

schema_name => 'scott',

streams_name => 'TEST_to_STANDBY',

source_queue_name => 'strmadmin.streams_queue',

destination_queue_name => 'strmadmin.streams_queue@STANDBY.com',

include_dml => true,

include_ddl => true,

source_database => 'TEST.com',

inclusion_rule => true,

queue_to_queue => true);

END;

/

PL/SQL procedure successfully completed.

可以通过dba_propagations查看结果：

SQL> select PROPAGATION_NAME,SOURCE_QUEUE_NAME,DESTINATION_QUEUE_NAME,DESTINATION_DBLINK,STATUS from dba_propagation;

PROPAGATION_NAME SOURCE_QUEUE_NAME DESTINATION_QUEUE_NAME DESTINATION_DBL STATUS

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

STREAM1_TO_STREAM2 STREAMS_QUEUE STREAMS_QUEUE STANDBY.COM ENABLED

10. 在source: TEST上创建Capture进程：

SQL>

BEGIN

DBMS_STREAMS_ADM.ADD_SCHEMA_RULES(

schema_name => 'scott',

streams_type => 'capture',

streams_name => 'capture_TEST',

queue_name => 'strmadmin.streams_queue',

include_dml => true,

include_ddl => true,

inclusion_rule => true);

END;

/

PL/SQL procedure successfully completed.

可将hr改为scott

可以通过dba_capture查看：

SQL> select CAPTURE_NAME,QUEUE_NAME,START_SCN,STATUS,CAPTURE_TYPE from dba_capture;

CAPTURE_NAME QUEUE_NAME START_SCN STATUS CAPTURE_TY

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

CAPTURE_STREAM1 STREAMS_QUEUE 504733 DISABLED LOCAL

SQL> select * from ALL_CAPTURE_PREPARED_SCHEMAS;

SCHEMA_NAME TIMESTAMP SUPPLEME SUPPLEME SUPPLEME SUPPLEME

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

HR 12-JUN-08 IMPLICIT IMPLICIT IMPLICIT NO

11.将TEST上的scott schema数据导到STANDBY上：

在stream2上：

可先将scott用户删除 drop user scott cascade;

sqlplus strmadmin/strmadminpw

SQL> create public database link TEST.com connect to strmadmin identified by strmadminpw using 'TEST';

Database link created.

SQL> select * from global_name@TEST.com;

GLOBAL_NAME

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

TEST

导数据：

[oracle@localhost admin]$ impdp strmadmin/strmadminpw network_link=TEST.com schemas=scott

Import: Release 10.2.0.2.0 - Production on Thursday, 12 June, 2008 17:47:59

Copyright (c) 2003, 2005, Oracle. All rights reserved.

.....

......

Processing object type SCHEMA_EXPORT/TABLE/STATISTICS/TABLE_STATISTICS

Job "SYSTEM"."SYS_IMPORT_SCHEMA_01" successfully completed at 17:50:32

遇到的问题：在导数据时发现导完后两边的表不一致，后发现在STANDBY从库上有scott用户及相关联的表，可使用命令drop user scott cascade 删除scott用户，再进行导数据即可解决此问题。

12. 在TEST上设置STANDBY上scott schema的Instantiation SCN

sqlplus strmadmin/strmadminpw

SQL>

DECLARE

iscn NUMBER;

BEGIN

iscn := DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER();

DBMS_APPLY_ADM.SET_SCHEMA_INSTANTIATION_SCN@STANDBY.com(

source_schema_name => 'scott',

source_database_name => 'TEST.com',

instantiation_scn => iscn,

recursive => true);

END;

/

PL/SQL procedure successfully completed.

13.在STANDBY上创建apply进程apply_STANDBY：

SQL>

BEGIN

DBMS_STREAMS_ADM.ADD_SCHEMA_RULES(

schema_name => 'scott',

streams_type => 'apply',

streams_name => 'apply_STANDBY',

queue_name => 'strmadmin.streams_queue',

include_dml => true,

include_ddl => true,

source_database => 'TEST.com',

inclusion_rule => true);

END;

/

PL/SQL procedure successfully completed.

可以通过：

dba_apply

v$streams_apply_reader

v$streams_apply_coordinator

v$streams_apply_server

查看状态

SQL> select apply_name,queue_name,status from dba_apply;

APPLY_NAME QUEUE_NAME STATUS

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

APPLY_STREAM2 STREAMS_QUEUE DISABLED

==启动==

13.启动capture和apply:

13.1 STANDBY上启动 Apply Process

CONNECT strmadmin/strmadminpw

SQL>

BEGIN

DBMS_APPLY_ADM.SET_PARAMETER(

apply_name => 'apply_STANDBY',

parameter => 'disable_on_error',

value => 'n');

END;

/

PL/SQL procedure successfully completed.

SQL>

BEGIN

DBMS_APPLY_ADM.START_APPLY(

apply_name => 'apply_STANDBY');

END;

/

PL/SQL procedure successfully completed.

SQL> select apply_name,queue_name,status from dba_apply;

APPLY_NAME QUEUE_NAME STATUS

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

APPLY_STREAM2 STREAMS_QUEUE ENABLED

这时候alert log有：

Thu Jun 12 18:00:36 2008

Streams APPLY A001 started with pid=25, OS id=30819

Streams Apply Reader started P000 with pid=26 OS id=30821

Streams Apply Server started P001 with pid=27 OS id=30823

如果有问题，没有能启动就查dba_apply的error_messages列

13.2 TEST上启动capture process:

sqlplus strmadmin/strmadminpw

SQL>

BEGIN

DBMS_CAPTURE_ADM.START_CAPTURE(

capture_name => 'capture_TEST');

END;

/

PL/SQL procedure successfully completed.

SQL> select capture_name,status from dba_capture;

CAPTURE_NAME STATUS

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

CAPTURE_STREAM1 ENABLED

alert 日志有：

Thu Jun 12 18:04:46 2008

Streams CAPTURE C001 started with pid=27, OS id=11884

Thu Jun 12 18:04:49 2008

LOGMINER: Parameters summary for session# = 1

LOGMINER: Number of processes = 3, Transaction Chunk Size = 1

LOGMINER: Memory Size = 10M, Checkpoint interval = 10M

LOGMINER: session# = 1, builder process P001 started with pid=32 OS id=11966

LOGMINER: session# = 1, reader process P000 started with pid=29 OS id=11964

LOGMINER: session# = 1, preparer process P002 started with pid=33 OS id=11968

Thu Jun 12 18:04:53 2008

LOGMINER: Begin mining logfile: /u01/oradata/stream1_arch/1_13_657197948.dbf

Thu Jun 12 18:06:40 2008

Thread 1 advanced to log sequence 15

Current log# 2 seq# 15 mem# 0: /u01/oradata/stream1/redo02.log

Thu Jun 12 18:06:53 2008

Thread 1 advanced to log sequence 16

Current log# 3 seq# 16 mem# 0: /u01/oradata/stream1/redo03.log

...

LOGMINER: Begin mining logfile: /u01/oradata/stream1_arch/1_13_657197948.dbf

Thu Jun 12 18:07:34 2008

LOGMINER: End mining logfile: /u01/oradata/stream1_arch/1_13_657197948.dbf

Thu Jun 12 18:07:34 2008

LOGMINER: Begin mining logfile: /u01/oradata/stream1_arch/1_14_657197948.dbf

信息，说明已经开始mining logfile.

===测试===

14.测试

在TEST上：

SQL> conn scott/tiger

select * from scott.emp;

update scott.emp set sal=sal+1 where empno=7369;

SQL> commit;

Commit complete.

STANDBY上检查结果(可能有延迟)：

SQL> conn scott/tiger

select * from scott.emp;

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

197 3001

这个时间间隔跟性能及其它情况有关，包括

capture mining logfile的速度

propagation的间隔(默认3秒)及传输时间

apply log的速度

问题诊断

5.1 如何知道捕捉(Capture)进程是否运行正常？

以strmadmin身份，登录主数据库，执行如下语句：

SQL>

SELECT CAPTURE_NAME, QUEUE_NAME, RULE_SET_NAME,NEGATIVE_RULE_SET_NAME,STATUS FROM DBA_CAPTURE;

结果显示如下：

CAPTURE_NAME QUEUE_NAME

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

RULE_SET_NAME NEGATIVE_RULE_SET_NAME STATUS

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

CAPTURE_PROD PROD_QUEUE

RULESET$_14 ENABLED

ENABLED

如果STATUS状态是ENABLED，表示Capture进程运行正常；

如果STATUS状态是DISABLED，表示Capture进程处于停止状态，只需重新启动即可；

如果STATUS状态是ABORTED，表示Capture进程非正常停止，查询相应的ERROR_NUMBER、ERROR_MESSAGE列可以得到详细的信息；同时，Oracle会在跟踪文件中记录该信息。

5.2 如何知道Captured LCR是否有传播GAP？

以strmadmin身份，登录主数据库，执行如下语句：

SQL> SELECT CAPTURE_NAME, QUEUE_NAME, STATUS, CAPTURED_SCN, APPLIED_SCN

2 FROM DBA_CAPTURE;

结果显示如下：

CAPTURE_NAME QUEUE_NAME STATUS

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

CAPTURED_SCN APPLIED_SCN

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

CAPTURE_PROD PROD_QUEUE ENABLED

17023672 17023672

如果APPLIED_SCN小于CAPTURED_SCN，则表示在主数据库一端，要么LCR没有被dequeue，要么Propagation进程尚未传播到从数据库一端。

5.3 如何知道Appy进程是否运行正常？

以strmadmin身份，登录从数据库，执行如下语句：

SQL> SELECT apply_name, apply_captured, status FROM dba_apply;

结果显示如下：

APPLY_NAME APPLY_ STATUS

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

APPLY_H10G YES ENABLED

如果STATUS状态是ENABLED，表示Apply进程运行正常；

如果STATUS状态是DISABLED，表示Apply进程处于停止状态，只需重新启动即可；

如果STATUS状态是ABORTED，表示Apply进程非正常停止，查询相应的ERROR_NUMBER、ERROR_MESSAGE列可以得到详细的信息；同时，可以查询DBA_APPLY_ERROR视图，了解详细的Apply错误信息。