So, Boost pumps primary job is to pressurize the fuel system to assist in engine start. In your scenario your attempting to start the #1 and #2 engines in an abnormal start, using just the DC Pump.
In my tests I performed the abnormal start you describe and I was able to successfully start both engines, and yes closing the crossfeed did shut down the #2 engine.
But I also did the test in a normal configuration. I then opened the crossfeed and I then secured both Right wing boost pumps. I got my normal warning indications. I then isolated the tanks by closing the crossfeed, and the #2 engine did operate normal.
To see if I would get any starvation, I then went to full power and had no starvation.
I wanted to test this high power test again since I had switched the boost pumps on/off and I reset the sim.
I did your abnormal start, and I tested the starvation of #2 and I did get the shutdown. I restarted the #2 engine and this time never turning on any boost pumps I went full power I got a #2 engine shut down , then a #1 engine shut down.
So, if this is a bug it would associated with never turning on the boost pumps. In all the tests that I did apply pressure to the #2 with boost pumps, after isolation with the crossfeed #2 engine did continue to operate.
I understand fuels systems, I am an A&P, but I am not an A-10 Mech, so without proper documentation I cannot say what the performance of losing both boost pumps in one tank would be. The engine fuel pumps are (should be) more than capable to supply fuel to the engine.
If I were to actually start these engines on a real A-10, I would expect them to be sufficient in fuel supply / pressure after they were brought online.
So I support it being called a bug, but in a normal start you would had turned on your boost pumps and my minimal testing showed that it should work normal in that case.