Instead of using a .ini file to hold the server name, I use a .udl file to define the server connection then pass the path to the .udl file as a parameter when running the DTS package. That way I don't need to mess around with environment variables, something that I may not have access to do on the production server. If you use SQL Server security, the user name & password are also stored in the .udl file so you don't need to worry about them in DTS.

Also, in an article like this you should have mentioned that table names in Data Pump tasks are recorded as [database].[owner].[tablename]. If you want those tasks to be database-independent you need to use a Dynamic Properties task to change them to [owner].[tablename].

Don