Yesterday we hit the dreaded conflict in our production/DR SQL peer-to-peer replication setup. My first reaction is usually “oh <expletive>, we’re going to need to rebuild replication”, but we happened to have the best possible scenario where we had an update-update conflict.
I’d like to note that I personally hate working with SQL Server peer-to-peer replication (we’re running SQL 2008 R2 version), and it looks like some of the limitations of the solution are still an issue in SQL 2012.
- We have to quiesce (i.e. stop the application) in order to add new articles (tables, views etc), to the replication topology. Big pain + off hours work.
- Plus, unlike with MySQL replication when setup in multi-master, I can’t just not log my correction statements so they don’t replicate back.
- This is why I’m excited for the Always On Availability groups in SQL 2012/2014, with a multi-subnet setup. I’m still evaluating our environment options but this helps address some of the limitations in our peer-to-peer replication environment.