Non-differentiable solution of the Brachistochrone problem

Question

Is there a solution to the brachistochrone problem where the solution is non-differentiable everywhere (angular point)? The Euler-Lagrange method fails if the first or second derivative of the solution do not exist.