Tensor Layout assemble c++

Question

I am trying to assemble vector just for 1 processor and I am running into seq. fault.

 TensorLayout local_vec_layout(MPI_COMM_SELF, {}, 0, 1, {}, false);
 dolfin::PETScVector local_vec;

   local_vec.init(local_vec_layout);
   assemble(local_vec, F); 

Any ides on bug?