Good question. There is another server in this whole configuration, let's call it Server 1. That's what the end users are hitting and updating. Server A copies data from key tables in Server 1 to the WarehouseDB in Server A and then the data is massaged on Server A and ultimately ends up on Server B. Transaction log is backed up on Server 1 every 3 hours. Web front ends hit Server 1 and Server B, but not A. Server A sits idly by until the next nightly processing window. I know... this is all rather confusing.