Hi @parachrosis! Sorry to hear things aren't going well 
. Unfortunately, we see this dada2 -9 exit code error happen when a machine runs out of memory. The easiest solution is to run this somewhere with more memory --- if you are on an academic cluster you could request more RAM. If not, you could try a QIIME 2 AWS EC2 instance, and run just this step there. Your final option is to process your samples as subsets, and then merge them after denoising. This isn't recommended in DADA2, but:
So, your mileage may vary. Good luck and keep us posted! 