Erik,
Yes, the results of your test seem odd. Sounds to me like one or both of the rails isn't perfectly straight. Not that odd, though. I've searched for, but never found, a manufacturing tolerance for straightness. I also seem to recall one vendor commenting on its website that the rails may not be perfectly straight 'out of the box,' so to speak. I believe that is why having a milled reference edge is preferred for rail mounting. You can clamp the rail to the edge and tighten it down. I used custom jigs and a precision straight edge when I set mine, but I had a relatively flat surface to work from. Apparently, you don't.
I do not believe anything you can make on a 3d printer is going to be up to the task of making an adequate linear rail surface. IMO, printed parts may be fine for making prototypes for testing, but not for a solution. I can't imagine plastic being even close to rigid enough. Also, not much strength for screw holding and threads holding up under vibration. However, making prototypes for testing purposes seems like a perfect diagnostic approach and modeling for a permanent solution.
Here's a reference source that includes information for mounting linear rails. You may find it useful. https://www.hiwin.com/pdf/linear_guideways.pdf
Gary