Pyspark Dataframe number of rows too large, how to avoid failure trying to count()?

Question

Trying to run some spark jobs.. when df.count() gets called, I get the following stack trace:

Starting the job...

Starting job: count at N