A unit disk with the Euclidean metric should be complete as a metric space.
But it is not geodesically complete I guess. Since every line will reach the boundary and can not be defined for all $t$.
But the Hopf–Rinow theorem says these two concepts are equivalent.
What is wrong in the above arguments...