Back in February I wrote about my cow orker Rohan, and his taunting me about his SETI@home stats.
Well, I let it get to me, and so being a semi-long weekend, I started running SETI@home on my Linux box on Friday. Only one instance, though, just to see what it would do. I noted that it seemed to be migrating from one CPU to the other, so I was hoping that it would keep the CPUs from overheating. I assume that overheating is why the machine craps out when running SETI@home, in spite of the overzealous cooling, although it could be something as simple as a scheduling bug in Linux.
Anyway, I let SETI@home run, and last night, after 133 days of uptime, and 5 days of running SETI@home, it crashed. And at first it wouldn’t boot because the BIOS couldn’t recognize the hard drive. A few blasts with the canned air and rejumpering the drive from “Master” to “Cable Select” (a hint from a friend) fixed that. I’ll never know if it was the jumpering or the cooling effect of the canned air that fixed the problem, but it’s fixed and I’m not running SETI@home again.
I lost a linux box to distributed.net, in almost the exact same way.
I run Windoze XP and have to reboot every night. I go to bed after I have completely shut down the computer and cold booted it and it runs with seti@home going for 24 hours and then I cold reboot. If I forget to reboot XP gets flaky after about three days.