I've been trying to figure out the best way to set up replication for our Disaster Recovery. Currently we have five production databases replicating through five distributors to one data warehouse. We're looking to recreate the same set of servers with replicated
data for our Disaster Recovery, but I can't figure out the best approach. We are using Server 2003, SQL 2005, and transactional replication. Between our production and Disaster Recovery, we are sharing a 45Mb/sec tunnel.
Here are the scenarios I've come up with:
Scenario 1: Replicate the production databases to our Disaster Recovery databases using the original publications and distributors. Then using our Disaster Recovery distributors, create new publications from
the replicated databases and replicate again to our Disaster Recovery Data Warehouse. With this method I've encountered a couple problems already. First, our two busiest databases can not keep up over the WAN after the snapshot is delivered (i.e. latency continues
to grow). and second, so far I haven't been able to replicate a replicated database, and I'm not even sure it is possible.
Scenario 2: Use the original production databases and configure our Disaster Recovery distributors to replicate to both our Disaster Recovery databases and our Disaster Recovery Data Warehouse. This means adding
a second distributor for each original publisher, and
View Complete Post