Hi All,
I have following situation:
NYC office has Oracle and SqlServer databases. NJ office has Oracle database.
Currently replication is going on between two oracle instance and it's healthy. **Both oracle instances are in the "Full" log mode.**
We plan to move NYC Oracle instance to MSSqlServe 2014. Now I need to replicate about 40+ tables to NJ Oracle instance from NYC MSSqlServe 2014 database using **transactional replication**.
Initial setup revealed replication was breaking after sometime due to timeout issues. We are past that issue now.
My real problem is the speed and size of some tables. I have about 15 tables with more than 1M rows. I did some basic benchmarking and realize the rate of propogation is about 6000 rows per minute. This will take more than 9 hours just for 1 table.
**The above speed is on a test oracle instance which is running in non-log mode.** I believe removing indexes on the Oracle (receiver side) could help but I also think doing that invalidates the replication and it begins all over again.
What could I do to improve this speed?
How you pros out there measure and keep tab on the process?
Any queries that you guys use for overall replication management?
Thanks!
DP
I have following situation:
NYC office has Oracle and SqlServer databases. NJ office has Oracle database.
Currently replication is going on between two oracle instance and it's healthy. **Both oracle instances are in the "Full" log mode.**
We plan to move NYC Oracle instance to MSSqlServe 2014. Now I need to replicate about 40+ tables to NJ Oracle instance from NYC MSSqlServe 2014 database using **transactional replication**.
Initial setup revealed replication was breaking after sometime due to timeout issues. We are past that issue now.
My real problem is the speed and size of some tables. I have about 15 tables with more than 1M rows. I did some basic benchmarking and realize the rate of propogation is about 6000 rows per minute. This will take more than 9 hours just for 1 table.
**The above speed is on a test oracle instance which is running in non-log mode.** I believe removing indexes on the Oracle (receiver side) could help but I also think doing that invalidates the replication and it begins all over again.
What could I do to improve this speed?
How you pros out there measure and keep tab on the process?
Any queries that you guys use for overall replication management?
Thanks!
DP