A counterexample on the oozing problem for closed manifolds