The answer is, as always, it depends...
A single nut is fine so long as (A) it is oriented such that work tightens it, and (B) that the rotational inertia (usually when stopping) is insufficient to loosen the nut. That's a bet I rarely take when I design machine tools! My personal preference when there is space is to make a slitted nut with a clamp-action screw. That makes it a torque to load and tighten clamp-screw operation. Jam nuts (with a washer in between the jam nut and the secondary nut) are my acceptable alternative choice.
The jam nut goes on the shaft first (after the interface packing and distribution washer). A washer (about .010 inches smaller in diameter than the wrench flats) made from something like aluminum (which has a very high coefficient of friction when mated to steel) provides a good fit. Then the full-size nut installed. Both nuts should be properly torqued (though I admit that the calibrated wrist works pretty well once the feel has been established).
I honestly have no idea how many shaft systems I have designed for grinders, wire-brush machines, or buffing machines over the years. I have never had a failure with either slitted/clamp-screw nut or jam nut/washer/nut assemblies using UNF threads. While it is preferred practice to use LH threads for the clockwise rotation side of a mounting shaft, I have (due to schedule & available tooling) built my fair share of such shaft system using RH threads with slitted/clamp-screw nut or jam nut/washer/nut retention that have worked quite well.
Does that help?
