Duplicate collision points and positive separation

Hi all

I have a strange issue, I get for example 30 collision points, however, some are duplicates, so it should be only about 15 which are unique. The remaining have the exact same position (not all are the same point, they are for example 5 different duplicate points).

And the second issue is, sometimes I get positive separation values. This should be the penetration depth, so what would this mean? Can I just ignore them?

I use PhysX 3.3.0 on Visual Studio.

Thanks for your help.