Planar defects like grain boundaries predetermine mechanical properties of polycrystalline materials and, in particular, their strength. Therefore, a lot of effort has been devoted to study these phenomena not only at macroscopic level but also at nano- and atomistic scales using simulations based on DFT. However, such simulations always omitted the Poisson contraction, which leads to relaxation of transverse stresses. In the present study, we propose and test two models of the tensile tests (including the transverse contraction) applied to crystals containing planar defects. One model comprises full optimization of the lattice via relaxation of the lateral stress tensor components while the other uses a new, simplified approach. The models are tested and verified for a tensile loading of the Σ5 (210) tilt grain boundary in Ni. The comparison of both methods reveals that the results are almost identical. However, our new model allows us to decrease the computational time significantly. Both models are also compared with former approaches that neglected the Poisson contraction.