OGG 10引入新token变量LOGCSN，对应于oracle的scn，对此replicat可以指定开始csn

Start replicat atcsn | aftercsn

准备阶段

同步前需确保源库的事务都已提交，应反复检查gv$transaction；

对于无法终止的事务，从其起始时间执行extract

eg:

sql> select min(start_time) from v$transaction/gv$transaction;

MIN(START_TIME)

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

02/08/13 19:21:38

The real-time extract can then be added by

ggsci> add extract tranlog begin 2013-02-08 19:20:00 --&gt> in case of standalone database

ggsci> add extract tranlog threads(number_of_threads) begin 2013-02-08 19:20:00 --&gt> in case of RAC

RMAN

假定source/target具有相同目录结构

--source

为source db添加supplemental日志

SQLPLUS > alter database add supplemental log data

Enable Supplemental Logging at Table Level

GGSCI> dblogin userid xxxxx password xxxxxx

GGSCI> add trandata .

添加extract/datapump进程和trail文件

备份数据库并查看其SCN

RMAN> backup database plus archivelog ;

RMAN> restore database preview summary;

Media recovery start SCN is 34821758

Recovery must be done beyond SCN 34822578 to clear datafile fuzziness

Finished restore at 10-OCT-10

将参数文件传送到target

--target

Startup nomount

使用duplicate命令clone数据库

OS> RMAN target sys/@

RMAN> connect auxiliary /

RMAN> duplicate target database to nofilenamecheck until scn ;

执行完毕后查看alert log

RESETLOGS after incomplete recovery UNTIL CHANGE 34822578

而SCN 34822578之后的工作则由replicat执行

启动replicat

GGSCI> add replicat , exttrail ./dirdat/

GGSCI> start replicat , atcsn

注：对于RAC系统，以上步骤大同小异；

唯一需要注意的是恢复RAC期间需要cluster_database设为false

228257.1

数据泵

--source

增添supplemental log

alter database add supplemental log data ;

Enable Supplemental Logging at Table Level on source

GGSCI> dblogin userid xxxxx password xxxxxx

GGSCI> add trandata .

获取current_scn

SQLPLUS> select current_scn from v$database ;

CURRENT_SCN

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

28318029

expdp directory=dumpdir full=y parallel=4 dumpfile=ora102_%u.dmp flashback_scn=28318029

--target

Impdp导入

GGSCI> add replicat , exttrail ./dirdat/

GGSCI> start replicat , aftercsn 28318029