What is the reasoning behind exporting the data to csv then importing it again?
Essentially the method is the same as the export just reversed, so you would have a connection manager to the file and a connection manager to serverB with a data flow task which takes data from the file and puts it into the database.