Comment (by Samuel Cupp):

If I make the change in IllinoisGRMHD’s schedule.ccl and run the test parfile magnetizedTOV.par, I don’t get any issues with stress_energy_at_RHS on or off. As such, I’m not sure if the issue is with IllinoisGRMHD itself. Particularly, the values that are nan when Con2Prim starts are the B field and the metric. The B field I could envision potentially having an issue if the setup for the scheduling were wrong, but the fact that all the spacetime quantities are nan seems to suggest there could be something wrong with the spacetime evolution or ID.

‌

--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2497/illinoisgrmhd-is-incompatible-with-setting