When ERESX, NO is used, the nodal solution is reported to be the same as the value at the integration point nearest to the node. Usually nodal values at 'extrapolated' from the values at the integration points.
I am not sure why that command worked once and is not working a second time.
But the stress values seen in the figure are way higher than the yield stress (almost 10 times). Irrespective of whether the nodal values are reported to be same as the values at the integration points or they are extrapolated from the integration points, I don't think the equivalent stress in your model will go below 68 MPa.
Is there a specific reason why you don't want the stress in the model to exceed the material's yield strength? Maybe the applied compression is so high that it actually generates high stresses. I am not sure if there is a plasticity model included to model actual yielding (if there is only an elastic model defined for the material, then the yield strength defined is used only to calculate the Factor of Safety during postprocessing; simply having a yield strength defined does not automatically model plastic behavior).