Dask crashing during shuffle (sort) on large dataframe with df.set_index?

Question

I\'m trying to call df.set_index on a large dataframe and am seeing lots of errors, followed by Dask worker crashes and restarts. Here is my setup:

Ubunt