March 24, 2003 at 10:49 am
I have just been introduced to a situation where someone proposes to set up replication in a chain topology. With 4 servers involved, Call them A, B, C, and D, they want to use transactional replication from A to C, and From B to D, and then merge replication between C and D, like this:
A.....B
|.....|
(T)...(T)
|.....|
C-(M)-D
I just cringe at the idea of this, but it really needs to be done, according to the business needs. Does anyone currently run a similar topology of replication successfully? How did they overcome the inherent issues with this?
Edited by - scorpion_66 on 03/24/2003 10:51:23 AM
March 24, 2003 at 12:04 pm
Does all kind of replication involves the same articles, I mean the same tables?
Because publisher A will undone any changes that the subscriber D made and that will be replicated to C with Merge replication.
March 24, 2003 at 1:01 pm
Yes, the tables replicated are the same. However, the data will be partitioned by data ranges within the tables (In a reliable manner I pray....).
What you mention is also one of the reasons it makes me cringe.....and I am looking for anyone who has seen anything like this work....
March 24, 2003 at 1:34 pm
I never use a situation like yours, the only problem I can think of is the one I told you. Can you try instead of using merge replication, using transactional replication with updating between server A to D?
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply