Seems fine to me, though I do not really understand the symmetry, since you fix everything.
But just for testing. If you substitute your remote displacement wiht a simple fixed constraint on that cylindrical surface, hopefully the problem disappears? It would be a first hint.