Generally you should only ground shielding at one point. If there is a good ground for the shield at the XLR then good practice says you shouldnt ground it at the star point either.

However the issue is minor here as there is little likelyhood of induced current in that short a run. The real issue is where you run a long shielded cable alongside, say, an unshielded power cable. Grounding both ends of the shield allows the magnetic field from the power cable to induce a current in the shield which can cause strange effects. In one site I worked on this problem caused the 'ground' of the remote device a few metres away to be 7v AC relative to the controller. With no opto isolation it wasn't surprising that the logic circuits were having some difficulty. Disconnecting the shield reduced the offset to <0.1V and all was well...