The RTC thing was a long shot It wouldn't have explained the incorrect updates, the clock doesn't get involved in the decision as to whether an aircraft gets updated or not (unless you have updates turned on, which you don't). A big leap in the real world clock could cause some temporary weird effects though, there are things that happen if things aren't updated for 10+ seconds etc. so a leap of several days could trigger those effects. I can't imagine how they would affect the online lookup caching though.
If I could get a debugger running on the Pi, and if it's as easy to reproduce on the Pi as you say, then I should be able to see what's happening fairly easily. Unfortunately I don't think it's easy to get a debugger running on the Pi. However I do have a few Pis doing things here, including one running VRS (although I don't want to mess about with that one). I might try putting some trace statements into a test build and see if they can shed some light as to what's happening, whereabouts things are going wrong. Also see whether I can reproduce it as well. My attempts so far have been on Windows. I didn't get a chance to do anything on VRS this weekend but I'll see if I can try it during the week.
Is Virtual Radar Server not behaving itself? If so then please report it here.