We are currently using replication to copy data from an OLTP database to a data warehouse. The problem is, whenever the application developers make a change to the db schema in Entity Framework, we have to take the affected objects out of replication, let the application deployment make the changes to the schema and then place the objects back into replication. We want to automate this process through Octopus Deploy, but scripting replication seems to be a real pain in the ... I looked into scripting the replication steps and it seems nearly impossible to do that. So I figured I would write an SSIS package to replace replication, create a SQL job to execute the package and then script the disable/enable of the package around the deployment step.
I've run into a bit of a problem however. The last time the package ran (in staging, _not_ in production, we still have replication running in production) it ran for just a little over 26 hours. It completed and everything moved around just fine, but 26 hours is far too long. So, I rewrote 4 of the biggest offending tables so that they use stored procedures for the inserts and updates and are now pushing those changes to staging in hopes that it will speed up the package. I was having SSIS performing all the checks, inserts and updates, but now the stored procedures are doing that.
Is there a better, faster way of doing this? Or, am I on the right track with the stored procedures?