Hi @Lisa_Crummett:
Real-world data-sets can take a bit of time on this step. Fortunately, once it is done a lot of the down-stream steps are pretty quick, relatively speaking!
Keep on waiting until it finishes or crashes 
There are no hard and fast rules here, but @benjjneb (the author of DADA2) has put together a bit of a case study, and also has this to say:
There are too many factors that influence running time to give an answer that will hold for everyone.
While that quote is referencing time, I think the same also applies (to some degree) to necessary computational resources.
If this fails, it likely means you don't have enough memory to process this dataset on the machine you are attempting to work on. Please see this post on a similar low-memory situation that has some suggestions on moving forward (specifically an institutional cluster, or AWS EC2). Sorry I don't have better news at this point. Thanks!