This seems to be a contentious area with views in different camps but, provided the endpoint is connected over Ethernet and the server meets the recommended specs then, for a machine running ROCK I think Roon staff have said there is no difference to audio quality. A server needn’t be expensive (no connection with seller).
Edit: FWIW I use the same spec server as linked and it is more than capable of doing all the DSP/upsampling I need. I also use the Lumin D2 you mentioned and have been very happy with it.