oracle数据库基于(streams流复制)的双机热备配置手册

-------------------------------------------------------------------------------
主数据库：
　　操作系统：windows 2012
　　IP地址：192.168.160.122
　　数据库：Oracle 11.2.0.2.0
　　ORACLE_SID：orcl
　　Global_name：orcl

　　从数据库：
　　操作系统：windows 2012
　　IP地址：192.168.160.123
　　数据库：Oracle 11.2.0.2.0
　　ORACLE_SID：orcl
　　Global_name：orcl

alter system set deferred_segment_creation=false scope=both;
alter system set sec_case_sensitive_logon=false scope=both;
ALTER PROFILE DEFAULT LIMIT PASSWORD_LIFE_TIME UNLIMITED;

主数据库配置
archive log list;
select supplemental_log_data_min from v$database;
shutdown immediate;
startup mount;
alter database archivelog;
alter database add supplemental log data;
archive log list;
select supplemental_log_data_min from v$database;
alter database open;

设置参数
alter system set aq_tm_processes=2 scope=both;
　　alter system set global_names=true scope=both;
　　alter system set job_queue_processes=10 scope=both;
　　alter system set parallel_max_servers=20 scope=both;
　　alter system set undo_retention=3600 scope=both;
　　alter system set nls_date_format='YYYY-MM-DD HH24:MI:SS' scope=spfile;
　　alter system set streams_pool_size=100M scope=spfile;
　　alter system set utl_file_dir='*' scope=spfile;
　　alter system set open_links=4 scope=spfile;
alter system set statistics_level='TYPICAL' scope=both;
alter system set utl_file_dir='*' scope=spfile;
alter system set "_job_queue_interval"=1 scope=spfile;
alter system set global_names=false scope=both;

shutdown immediate;
startup;

create user hytera identified by hytera account unlock;
grant connect to hytera;
grant RESOURCE to hytera;
grant dba to hytera;

--alter system set logmnr_max_persistent_sessions=1 scope=spfile;

create tablespace streams_tbs datafile 'D:\app\Administrator\oradata\orcl\streams_tbs.dbf' size 100M autoextend on maxsize unlimited;

create user strmadmin identified by strmadmin default tablespace streams_tbs account unlock quota unlimited on streams_tbs;
grant dba to strmadmin;
grant connect to strmadmin;
exec dbms_streams_auth.grant_admin_privilege('strmadmin');

begin
dbms_streams_auth.grant_admin_privilege(
grantee=>'strmadmin',
grant_privileges=>TRUE);
end;
/

tnsnames.ora
SZXTDB =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.160.123)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = orcl)
)
)

conn strmadmin/strmadmin;
create database link SZXTDB connect to strmadmin identified by strmadmin using 'SZXTDB';

创建Master流队列
#以strmadmin身份，登录主数据库。
connect strmadmin/strmadmin;
begin
dbms_streams_adm.set_up_queue(
queue_table => 'source_queue_table',--队列表
queue_name => 'source_queue',--队列
queue_user =>'strmadmin'); --队列用户
end;
/
select owner,queue_table,name from dba_queues where owner='STRMADMIN';

创建数据捕获进程,捕获redo的改变

connect strmadmin/strmadmin;
begin
dbms_streams_adm.add_schema_rules(
schema_name => 'hytera', --捕获用户名 非sys或者system
streams_type => 'capture',
streams_name => 'capture_stream',
queue_name => 'strmadmin.source_queue',
include_dml => true,
include_ddl => true,
include_tagged_lcr => false,
source_database => 'ORCL',
inclusion_rule => true);
end;
/

select capture_name,queue_name,start_scn,status,capture_type from dba_capture;

在源数据库上创建传播进程 相当于ogg里面的datapump数据传递
传播进程需要对应捕获规则分别创建table/schema/global的传播进程。
connect strmadmin/strmadmin;
begin
dbms_streams_adm.add_schema_propagation_rules(
schema_name => 'hytera',
streams_name => 'source_to_target',
source_queue_name => 'strmadmin.SOURCE_QUEUE',
destination_queue_name => 'strmadmin.TARGET_QUEUE@szxtdb',
include_dml => true,
include_ddl => true,
source_database => 'ORCL',
inclusion_rule => true,
queue_to_queue => true);
end;
/

select propagation_name,status from all_propagation;

直接设置SCN的方式进行实例化
查询源端的scn号
select DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER() FROM dual;

启动捕获进程 相当于OGG里面的ext1
BEGIN
DBMS_CAPTURE_ADM.START_CAPTURE(
capture_name => 'capture_stream');
END;
/

停止捕获进程
BEGIN
DBMS_CAPTURE_ADM.STOP_CAPTURE(
capture_name => 'capture_stream');
END;
/

查看捕获进程状态
select capture_name,status from dba_capture;

表示实时传播lcr实时复制
begin
dbms_aqadm.alter_propagation_schedule(
queue_name => 'strmadmin.SOURCE_QUEUE',
destination => 'szxtdb',
destination_queue => 'TARGET_QUEUE',
latency => 0);
end;
/
-------------------------------------------------------------------------------------------------------------------------
备数据库配置
archive log list;
select supplemental_log_data_min from v$database;
shutdown immediate;
startup mount;
alter database archivelog;
alter database add supplemental log data;
archive log list;
select supplemental_log_data_min from v$database;
alter database open;

设置参数
alter system set aq_tm_processes=2 scope=both;
　　alter system set global_names=true scope=both;
　　alter system set job_queue_processes=10 scope=both;
　　alter system set parallel_max_servers=20 scope=both;
　　alter system set undo_retention=3600 scope=both;
　　alter system set nls_date_format='YYYY-MM-DD HH24:MI:SS' scope=spfile;
　　alter system set streams_pool_size=100M scope=spfile;
　　alter system set utl_file_dir='*' scope=spfile;
　　alter system set open_links=4 scope=spfile;
alter system set statistics_level='TYPICAL' scope=both;
alter system set utl_file_dir='*' scope=spfile;
alter system set "_job_queue_interval"=1 scope=spfile;
alter system set global_names=false scope=both;

shutdown immediate;
startup;

--alter system set logmnr_max_persistent_sessions=1 scope=spfile;

create user hytera identified by hytera account unlock;
grant connect to hytera;
grant RESOURCE to hytera;
grant dba to hytera;

create tablespace streams_tbs datafile 'D:\app\Administrator\oradata\orcl\streams_tbs.dbf' size 100M autoextend on maxsize unlimited;

create user strmadmin identified by strmadmin default tablespace streams_tbs account unlock quota unlimited on streams_tbs;
grant dba to strmadmin;
grant connect to strmadmin;
exec dbms_streams_auth.grant_admin_privilege('strmadmin');

begin
dbms_streams_auth.grant_admin_privilege(
grantee=>'strmadmin',
grant_privileges=>TRUE);
end;
/

tnsnames.ora
SZXTDB =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.160.122)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = orcl)
)
)

conn strmadmin/strmadmin;
create database link SZXTDB connect to strmadmin identified by strmadmin using 'SZXTDB';

创建备库流队列
#以strmadmin身份，登录主数据库。
connect strmadmin/strmadmin;
begin
dbms_streams_adm.set_up_queue(
queue_table => 'target_queue_table',--队列表
queue_name => 'target_queue',--队列
queue_user =>'strmadmin'); --队列用户
end;
/
select owner,queue_table,name from dba_queues where owner='STRMADMIN';

目标端数据库上创建Apply队列应用进程 相当于OGG中的rep1进程
connect strmadmin/strmadmin;
begin
dbms_streams_adm.add_schema_rules(
schema_name => 'hytera', --捕获用户名 非sys或者system
streams_type => 'apply',
streams_name => 'target_apply_stream',
queue_name => 'strmadmin.TARGET_QUEUE',
include_dml => true,
include_ddl => true,
include_tagged_lcr => false,
source_database => 'ORCL',
inclusion_rule => true);
end;
/

select apply_name,queue_name,status from dba_apply;

设置成跟源端一致的SCN号
源端查
select DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER() FROM dual;
conn strmadmin/strmadmin;
BEGIN
DBMS_APPLY_ADM.SET_SCHEMA_INSTANTIATION_SCN(
source_schema_name => 'hytera',
source_database_name => 'ORCL',
instantiation_scn => 1097715);
END;
/

启动目的端的Apply应用进程相当于启动OGG里面的rep1
BEGIN
DBMS_APPLY_ADM.START_APPLY(
apply_name => 'target_apply_stream');
END;
/

停止apply
BEGIN
DBMS_APPLY_ADM.STOP_APPLY(
apply_name => 'target_apply_stream');
END;
/

查看apply进程状态
select apply_name,queue_name,status from dba_apply;
--------------------------------------------------------------

查看捕获进程状态
select capture_name,status from dba_capture;

查看apply进程状态
select apply_name,queue_name,status from dba_apply;

在源数据库上创建传播进程 相当于ogg里面的datapump数据传递
select propagation_name,status from all_propagation;

select capture_name,status from dba_capture;
select apply_name,queue_name,status from dba_apply;
select capture_name,status from dba_capture;
select error_message from DBA_APPLY_ERROR;

源端查询 抓取进程和应用进程的SCN号是否一致,流复制是基于redolog 的scn号的
select capture_name,queue_name,status,captured_scn,applied_scn from dba_capture;

查看应用进程是否有错误
select apply_name,error_number,apply_captured,status from dba_apply;

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

启动捕获进程 相当于OGG里面的ext1
BEGIN
DBMS_CAPTURE_ADM.START_CAPTURE(
capture_name => 'capture_stream');
END;
/

停止捕获进程
BEGIN
DBMS_CAPTURE_ADM.STOP_CAPTURE(
capture_name => 'capture_stream');
END;
/

启动目的端的Apply应用进程相当于启动OGG里面的rep1
BEGIN
DBMS_APPLY_ADM.START_APPLY(
apply_name => 'target_apply_stream');
END;
/

停止apply
BEGIN
DBMS_APPLY_ADM.STOP_APPLY(
apply_name => 'target_apply_stream');
END;
/

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

create tablespace hytera datafile 'D:\app\Administrator\oradata\orcl\hytera.dbf' size 100M autoextend on maxsize unlimited;
alter user hytera default tablespace hytera;

--------------------------------------------------------------------------------
配置双向流复制

源端执行
解决ORA-26687问题

SELECT * FROM DBA_CAPTURE_PREPARED_TABLES;

BEGIN
DBMS_APPLY_ADM.SET_TABLE_INSTANTIATION_SCN (
source_object_name => 'hytera.t',
source_database_name => 'ORCL',
instantiation_scn => 1094239);
END;
/

select error_message from DBA_APPLY_ERROR;--流复制错误信息

PURGE recyclebin;

解决 ORA-00942 问题
BEGIN
DBMS_APPLY_ADM.SET_PARAMETER(
apply_name => 'target_apply_stream',
parameter => 'disable_on_error',
value => 'n');
END;
/
--------------------------------------------------------------------------------

drop user hytera cascade;
create user hytera identified by hytera account unlock default tablespace hytera;
grant connect to hytera;
grant RESOURCE to hytera;
grant dba to hytera;