I've got a bearing contact surface with 10mm wide cylinder of 13mm radius against a flat surface.  (Assume steel on steel for starters.)  I'm getting no where near what the theoretical Hertzian stresses should be in a real model I'm working on.  When I run a simplified model and get the mesh down to the order of 1 mm then the stresses are about what they should be (~112ksi)  (Yes, I know I'm mixing units but that's not the problem.)

When I run the same on a full model I get totally unrealistic numbers on the order of 10-50 ksi depending on the mesh size. When I use local volumes at 1 mm mesh size I end up with about 25ksi.  Still totally unrealistic.  Linear Static gives bigger numbers than the non-linear static.

SW Premium 2014 x64 SP4