After following Konstantin's debugging suggestions, I thought I would try to replicate 
the issue on my own computer.  My hope was that I could provide instructions for 
replicating the bug so that the MIDAS team could try debugging things more easily.

However, when I ran the current version of MIDAS in a Rocky 9.4 VM on my laptop (both 
VMWare and VirtualBox), mserver and odbedit ran just fine (!).

I'm currently trying to find out if there's a way to compare the VMs on my machine and 
the machine that's being problematic, I'll report back if I learn anything.