Nodal stress and strain values (usually used only for output) can be wrong if mechanics and flow are coupled in a partitioned way. This is because nodal strain values are always summed up for each solving DEFORMATION process. However, in each coupling iteration, the values should be reset to previous time step values. Nodal displacement and gauss point stress/strain values are still correct.
UPDATE: it seems nodal stresses are correct.