Golden Gate version: 11.2.1.0.3
RDBMS version : 11.2.0.3
OS : RHEL 5.4
From Production OLTP (Source) to Production ETL (Target), we want to replicate 5 tables.
Each table is around 3 gb in size in Source now.
In Source , around 200 mb of data is being INSERTed every hour in all these 5 tables combined .
This is how we are planning to do the Iniitial Load:
After installing goldengate on both servers (and keeping the parameter files ready), we are going to take and exdp dump of these 5 tables from Source and SCP it to target server and import it there using impdp.
Total estimated time for expdp, scp the dumpfile and impdp in target DB is: 2 hours
During this gap of 2 hours , 400 mb worth of data would have been INSERTed in these 5 tables in the source. We want the tables to be in Sync in both Source and Target. For that ,
1. Shouldn't we be leaving the extract at the source running so that it will capture the changes during these 2 hours ?
2. What about the secondary extract (Pump) in source. Should it be running ?
3. Replicat in the Target should be stopped during this period of 2 hours. Right ?
Anything else we have to do to make this configuration smooth ?