| than the rest of the office (small company, all the servers are high end towers), but as soon as you set foot in the room you notice a big temp change.
But the 42C was also the temp once they got in to work on Tuesday after 12+ hours of no cooling in the server room, it could have been much warmer than that after the a/c was turned off, and after the heat tripped some/most of the thermal cutoffs the temp would drop back down (interior walls usually don't have any insulation).
I'm assuming your room is ~22C, so if your CPU idles at 52 it's a 30C increase... if you warmed your room up to 42C (like the server room was) that same difference would push your cpu up closer to 72C. Add loads and any higher room temp as I said above and the systems could easily be >90C. Once a single load-balanced system failed the extra load that has to be passed on the other systems would only speed up their failure as well. The article didn't say what failed, and I know CPU's or video cards can take a much higher temp than hard drives, some capacitors, etc.
OTOH, if it wasn't believable it wouldn't make much of a "true story", so it could go either way.
|
