How to order ETL tasks in Sql Server Data Tools (Integration Services)?

I'm a newbie in ETL processing. I am trying to populate a data mart through ETL and have hit a bump. I have 4 ETL tasks(Each task filling a particular table in the Mart) and the problem is that I need to perform them in a particular order so as to avoid constraint violations like Foreign Key constraints. How can I achieve this? Any help is really appreciated.

This is a snap of my current ETL:

Create a separate Data Flow Task for each table you're populating in the Control Flow, and then simply connect them together in the order you need them to run in. You should be able to just copy/paste the components from your current Data Flow to the new ones you create.

The connections between Tasks in the Control Flow are called Precendence Constraints, and if you double-click on one you'll see that they give you a number of options on how to control the flow of your ETL package. For now though, you'll probably be fine leaving it on the defaults - this will mean that each Data Flow Task will wait for the previous one to finish successfully. If one fails, the next one won't start and the package will fail.

If you want some tables to load in parallel, but then have some later tables wait for all of those to be finished, I would suggest adding a Sequence Container and putting the ones that need to load in parallel into it. Then connect from the Sequence Container to your next Data Flow Task(s) - or even from one Sequence Container to another. For instance, you might want one Sequence Container holding all of your Dimension loading processes, followed by another Sequence Container holding all of your Fact loading processes.

A common pattern goes a step further than using separate Data Flow Tasks. If you create a separate package for every table you're populating, you can then create a parent package, and use the Execute Package Task to call each of the child packages in the correct order. This is fantastic for reusability, and makes it easy for you to manually populate a single table when needed. It's also really nice when you're testing, as you don't need to keep disabling some Tasks or re-running the entire load when you want to test a single table. I'd suggest adopting this pattern early on so you don't have a lot of re-work to do later.