Check_totals fails when objective is an output of a parallel component

pargroup_output_as_objective.py

import numpy as np 
import openmdao.api as om 
import time


class DelayComp(om.ExplicitComponent):

    def initialize(self):
        self.options.declare('time', default=3.0)
        self.options.declare('size', default=1)

    def setup(self):
        size = self.options['size']
        self.add_input('x', shape=size)
        self.add_output('y', shape=size)
        self.declare_partials('y', 'x')

    def compute(self, inputs, outputs):
        waittime = self.options['time']
        if not inputs._under_complex_step:
            print('sleeping: ')
            time.sleep(waittime)
        outputs['y'] = inputs['x']

    def compute_partials(self, inputs, J):
        size = self.options['size']
        J['y', 'x'] = np.eye(size)



if __name__ == "__main__":
    model = om.Group()
    iv = om.IndepVarComp()
    size = 1
    iv.add_output('x', val=3.0 * np.ones((size, )))
    model.add_subsystem('iv', iv)
    pg = model.add_subsystem('pg', om.ParallelGroup(), promotes=['*'])
    pg.add_subsystem('dc1', DelayComp(size=size, time=2.0))
    pg.add_subsystem('dc2', DelayComp(size=size, time=2.0))
    pg.add_subsystem('dc3', DelayComp(size=size, time=2.0))
    model.connect('iv.x', ['dc1.x', 'dc2.x', 'dc3.x'])
    model.add_subsystem('adder', om.ExecComp('z = sum(y1)+sum(y2)+sum(y3)', y1={'value': np.zeros((size, ))},
                                                                            y2={'value': np.zeros((size, ))},
                                                                            y3={'value': np.zeros((size, ))}))
    model.connect('dc1.y', 'adder.y1')
    model.connect('dc2.y', 'adder.y2')
    model.connect('dc3.y', 'adder.y3')

    model.add_design_var('iv.x', lower=-1.0, upper=1.0)
    # this objective works fine
    # model.add_objective('adder.z')

    # this objective raises a concatenation error whether under fd or cs
    model.add_objective('dc1.y')

    # for some reason this constraint is fine even though only lives on proc 3
    model.add_constraint('dc3.y', lower=-1.0, upper=1.0)

    prob = om.Problem(model=model)
    prob.setup(force_alloc_complex=True)
    prob.run_model()
    # prob.model.list_outputs(print_arrays=True)
    prob.check_totals(method='cs', compact_print=True)