Peter Geoghegan <pg@heroku.com> writes:
> Even when the leader is consuming input from workers, that's still perhaps
> pegging one CPU core. So, it doesn't really invalidate what I said about
> the number of cores being the primary consideration.
Agreed, but if we think that people need to be thinking in those terms,
maybe the parameter should be "max_parallel_cores".
The alternate docs patch I just posted tries to deal with this by
describing max_parallel_workers as being the max number of worker
processes used to "assist" a parallel query. That was terminology
already being used in one place, but not consistently. If we use it
consistently, I think it would be sufficient to remind people that
they need to figure on one more core for the leader.
regards, tom lane
