Inspiration for this post comes from Gizmodo's hard drive failure song contest. Jess sent me the link a little while back and I've been following it. Winner's have been announced as you can see.
WTF am I talking about, you ask? Hitachi posted several sound files of failing hard drives to their support site. The idea was that you could diagnose your hardware failure by matching the noise.
My personal fave is this. Nice breaks. You could easily play this out in town and people would groove to it. Though this one, the winner, is pretty excellent in its stripped down simplicity and the fact that you're exclusively hearing the sound of bits dying. Add some creepy German accented vocals and you have Kraftwerk's first album. And how can you miss this one, featuring music so awful...its well...awful, but it's kinda funny.
This all reminds me of the first time I set computer equipment on fire. Back in 1995, I was the chief geek at my University library. It was really a pretty excellent student job as these things go. I got free reign to set things up how I liked and got to do all of the forward looking new projects. We were deploying a new electronic journal archive and I was getting the server all ready. For storage we had just bought a big external enclosure tower into which we wanted to put 4 new $1500 1.5gig drives in it. This gave us 6 gigabytes of storage, which my boss thought was overkill, but I had convinced him that we needed it. Anyway, today was the day to put it all together and I was psyched!
The first part was to mount each drive into the enclosure and get everything hooked up right. Then plug the tower into the server, boot the thing, and then try to convince SCO UNIX to recognize this puppy. None of this was supposed to be hard. So off I went, tools in hand and got it all set up and plugged in. Time to flip the power switch. No drama at this step, right? Gosh, Andy, why is smoke pouring out of the new enclosure? Moment of hesitation, as I worry for the filesystem and then realize that my drives are on fire and therefore a “graceful shutdown” isn’t the top priority. Kill the power. Deep breath. Fuck, it really really smells like burning plastic in here. I hope the smoke alarm doesn’t go off. Let me open the door (to one of the main floors of the library) and air the place out.
A few minutes pass while I sort myself out. I start to realize that the intense aroma of broiled electronics that had been centered in my office has now permeated my entire floor and is working it’s way to other ones. People are apparently complaining to circulation. My boss, Roger, was out at lunch, so now the library staff is calling me to see if everything is ok. I try to assure them that yes it’s ok and that no they don’t need to call any facilities people. I watch through the door as a couple of students pack their books up to presumably move to a lower emissions area.
Time to find out went wrong and how bad it is. I take the enclosure apart and pull out the drives. They’re on the floor when Roger walks in. “ANDY, what’s going on in here?” <he sees the drives> “THAT’S $6000 of BRAND NEW GEAR, WHAT THE HELL IS GOING ON?”. “Roger, can you give me a little while? I don’t really know.” He’s nice and leaves me alone for a bit. I should probably go outside, because frankly I’m feeling dizzy now, but instead I investigate the drives. A touch of context here, in this tower enclosure the drives all sit on one another in a neat stack. My neat stack now had one with it’s control board in a state best described as charred, complete with ashy black residue and a nice array of melted components. Clearly this was the source of the fire. The unit just above this one might have worked, but it was sticky and gooey just outside where the platters should be. I wasn’t hopeful.
Looking back more closely at the strike-anywhere drive, I can spot the problem. It seems somehow the rail mounting screw I used was too long and managed to short something so important that it actually started a fire. It’s sort of funny, because when you’re doing computer shit, after having to deal with so many brutal-to-debug intermittent problems, it’s almost kind of satisfying to have a problem so easy to spot like this. But anyway, look at that, Roger’s back and I get to explain what happened. I was a good boy and took the blame. He decided to be a good boy as well and announced that the problem wasn’t that I used the wrong screw (which of course, I had), but that the drive was poorly designed and he was fighting mad to kick some ass with this hardware vendor. So I got angry with him (augmented a bit I’m sure by my nice plastic fume carcinogen high) and we psyched him up to “show them a piece of his mind”.
Bafflingly the vendor did take the blame and replaced both drives. And I got invited to Roger’s house for dinner. We played ping pong in his garage. It was fun.