Microsoft server crash leaves 800 planes in the air without air traffic control | Hacker News Reader

Microsoft server crash leaves 800 planes in the air without air traffic control

techworld.com

18 pointskirubakaran15 comments

Open articleView on HN

The artcile is dated 21 September 2004. Does it make sense to revisit something 4 years old?

(I am new to hacker news, not sure if posting older articles are okay)

No, it doesn't make sense. But it's a cheap way to get karma points.

Not if the issue is still relevant :-)

"The servers are timed to shut down after 49.7 days of use in order to prevent a data overload"

Sounds like improperly designed software and it's not the first time I've seen people try to sell solutions that require a periodic reboot to cover up for poorly written code. Why in the world does a system in charge of air traffic control need to be rebooted every 2 months?

More importantly, sounds like MS is not really to blame.

I would have loved to say "that's what happens when you switch from UNIX to Windows servers" but you're right, it doesn't look like it's caused by Microsoft.

[49.7 days in milliseconds]

49.7 days = 4 294 080 000 milliseconds

[2^32]

2^32 = 4 294 967 296

It's not 'data overload' but overflow of an unsigned 32-bit int time field they're worried about.

it almost certainly has to do with the win32 function GetTickCount(), which rolls over every 49.7 days

http://en.wikipedia.org/wiki/GetTickCount

Didn't Windows 95 have a bug where it would crash after 49.7 days because the counter used to keep track of its uptime overflowed?

i don't think that's true. good luck getting a machine to stay up that long, but it's definitely possible.

Why would one want to use windows for something as critical as ATC? As if there weren't enough problems out there with antiquated planes, etc. Yikes!

During new year celebrations in Seattle, they had to reboot Windows right in the middle of the fireworks! That took too long , so they had to set off the pyro manually.

http://kirubakaran.blogspot.com/2008/01/new-year-gift-from-o...

That was no way of the same magnitude as the ATC crash, but still the failure was so real and viivd as I was there in person.

The blog you link to, http://blogs.king5.com/archives/2008/01/computer_glitch.html, says that the error was due to a corrupted file (assuming this means for the firework control software).

First, when questioned by the TV crews, the pyro guys refused to say what OS they were running. They did say that they had to reboot the computer. Then later (may be after suits with duffel bags arrived) they said that it was just due to a corrupted file. It was a screw up big enough to be worthy of a cover up (since it touched the common man).

It sounds like a nut-job conspiracy theory but I have difficulty giving MS the benefit of the doubt.

"The shutdown is intended to keep the system from becoming overloaded with data and potentially giving controllers wrong information about flights, according to a software analyst cited by the LA Times."

What the hell? Data overload?? Maybe someone should tell them about garbage collection?...

Daniel

Reply on news.ycombinator.com