|SQL Server to Oracle Replication - Golden gate - Checkpoint info and backup queries [message #617904]
||Fri, 04 July 2014 05:26
Registered: June 2014
Location: Dark Saturn
We are trying to replicate database from SQL Server 2008 to Oracle 11g R2 database (on linux) using Golden Gate Software 11g
In this regard I have read some wonderful articles but still I have the following queries
Could you please suggest on the following?
I understand we put the SQL Server database in 'Full' recovery mode before we start the 'Initial Load' and subsequent 'Continuous Replication'
a) Do we take backup of the source database, only before 'Initial Load' or we continue taking the backup as per our regular policy and it would not obstruct the Extract process to read from Tran Log which I believe is truncated after the backups?
b) If we are trying this in development environment, is it mandatory to take backups before and (that too) in regular intervals during the replication process?
(Assuming we are Not running out of space)
In other words does the replication has any dependency with the backups?
After the initial load and while starting the continuous replication do we need to configure the checkpoint information for the Extract?
Or the extract handles it implicitly?
I understand we disable foreign key constraints and all triggers (Non System) on the target side
a) Do we need any further configuration on the target side as far as triggers, constraints or indexes are concerned?
b) Do we need to configure these objects on the source database as well in any manner?
In worst case if the 'initial load' fails is there a way we clean up the target tables and
a) restart the initial load from the already available trail files on the target server ?
b) again restart the initial load, right from the start e.g.extract process?
Is it advantageous to use the (data)pump in the continuous replication over not using it at all?
Thank you so much in advance