I think I may have a solution for my pump cavitation problem. I've done some additional research and I agree that it appears to be a cavitation issue. I'm likely going to have to redesign the pump discharge to be a vertical system with a bleed valve. Currently it is a horizontal discharge and that may be causing the pump some difficulty in removing air bubbles. I'm still curious as to why the first fill didn't produce this problem.

I wonder if the glycol (being somewhat viscous) caused a vapor barrier behind the impeller that caused an internal air bubble that couldn't force its way out. Sort of the way soapy water forms a bubble. Once I filled, at static pressure it could have been enough to maintain itself in that tiny space. Once running the bubble couldn't be ejected. It might have settled at the top of the volute and stayed... Which is odd... Of course, that is all just a supposition... Either way, it seems to be cavitation.