Essentially you can query for all of the data you want to retrieve from the database on the source node, however, as a general rule, we like to run smaller queries to reduce the time and number of rows that may get locked while retrieving large amounts of data on the source node.
I like to query for an id and some basic patient identifying data on the source node so that I can search for messages by patient name or id on the message tab later on, and then in the mapping use subqueries to gather all of the other relevant data for a given patient/message.
Here is a channel that can be downloaded as an example.
Once you get the channel you can import it by following these instructions: https://www.qvera.com/kb/index.php/364 which also is how you can export and import a channel from your test server to your prod server.