I have been to Kalkar, the never-commissioned nuclear plant near the German/Dutch border, now a theme park.
All I can say is that the walls of the containment vessel are indeed substantial, and in fact there is a number of layers between the outside air and the reactor core.
Having said that, the incident in Japan showed us that you can have a serious nuclear accident even when the walls of the containment vessel are not breached at all. There is a lot of equipment located outside the containment wall, yet vital for the well-being of the plant.
I wonder if a significant fire, say resulting from a 9/11-like event, outside the containment wall and raging for, say, 24 hours, has ever been modeled. Not just on the structural integrity of the containment wall, but also on electricity feed for the cooling systems, the cooling systems themselves, the operators and so forth. After all, on 9/11 it was not the impact that brought the towers down, but the resulting fire.