OK, but wouldn't a force applied in the direction of the centripetal arrow in your diagram result in an angular acceleration? I am assuming that the cylinder is rotating at constant angular velocity so I can't see where this force comes from