Hello,
I have a few questions regarding SQL replication and how it handles large sets of data. We are exploring different options to offload reporting from the OLTP system. The database in question is almost 1TB in size. The tables that are used for reporting have millions and millions of records. Some have billions. We currently have a mirror setup and do a snapshot and then report off the snapshot. The business now requires almost "real-time" data for their reports instead of hourly snapshots. It's a daunting process for the server to keep generating a snapshot and the application has many errors from it because the server is constantly generating the snapshot with the latest data. Long story short the requirements have changed so we are looking into transactional replication for our reporting solution.
We'd like to keep the mirror in place and add in another server(s) for reporting. We are thinking two more SQL servers. One subscriber and one distributor. The existing servers would be the principal and mirror. From what I've been reading it's best to have the distributor database on a dedicated server.
- What would be the space requirements for the distribution database?
- Are there any potential issues with replicating such large sets of data?
Thank you for your time.