How do the clocks get back to a shared location if their motion is uniform? Within a special relativistic framework, there's no way.
Now if you include a general relativistic curved spacetime, it's possible that the overall geometry is closed and that the two clocks, continuing on inertial paths, will ultimately meet up again. That raises interesting questions as to whether in such a geometry there actually would be a globally preferred reference frame, even though local patches have no such preferred frame. (See https://journals.aps.org/prd/abstract/10.1103/PhysRevD.8.1662, https://journals.aps.org/prd/abstract/10.1103/PhysRevD.8.1662)
Mike W.
(published on 03/28/2018)