It Got Around To Me

The web server upgrade was grand for a while, but then I started getting mysterious, daily kernel panics. While the occasional glitch is expected for any machine built out of parts that are just lying around, 40% downtime does not indicate a respectable level of reliability, even for the $0 price. After all, having a slick’n'fast WordPress installation is worthless if it’s never actually running when I want to post!

The Oracle told me that it was probably bad RAM, but it ALWAYS tells me that.[1] I grant that sometimes the problem actually is bad RAM, and I accept that the problems caused by RAM in any state between “works almost perfectly” and “is almost DOA” can be hard to diagnose, and I concede that you should probably be running Memtest86+[2] on the crashed machine while trawling for the solution to your problem.

And, hey! There we go. Memtest has found some tests that Shall! Not! Pass! and it’s like Christmas on the 5″ server monitor with all the red and white blinking by on the blue background… with a blinky green thing in the corner… did these interface guys do this on purpose? Maybe it’s to remind me to ask Santa for new RAM… Try to focus, even though it’s after Midnight and the Woot-Off has been going on for 24 hours now…

So, of course, I test each stick-o-RAM individually to isolate the error and eliminate it. #1 is bad… #2 is bad, too, ouch… #3 is bad?… EVERY STICK OF RAM IN THE HOUSE IS BAD?!… No, no… I think the motherboard’s RAM controller has gone bad.

This is why, from 2:00 to 3:00 AM, I was in my server closet, wearing my jammies and slippers, rebuilding the server onto another old motherboard in another old case and testing it. And it works![3] But the new-old motherboard won’t support the 1.5GHz processor, so we’re now running on a 750MHz Duron. That’s okay, I guess, but for the cost of about 1GB of PC133 SDRAM, I could do so much better

TO BE CONTINUED (PROBABLY FOR THE REST OF MY LIFE)…

  1. I think you could probably look up the causes of bad breath, for instance, and if you get to any sort of technical forum or chat log, you’ll find exchanges like this one:

    Zeta109: I’m having some trouble with my breath. It always stinks. Maybe I should brush my teeth sometimes. Or should I stop putting diced onions on my Pop Tarts?

    TekyD00d: Yeah, you could try that stuff, but make sure to test your RAM thoroughly as well. Let the tests loop for a few days, in case it’s not particularly bad, and the errors need some coaxing to surface.

    UthaTek: Totally. Bad RAM even killed my cat, once. []

  2. I run it from a Knoppix CD, which is the height of convenience for a dude like me. []
  3. And hopefully, the motherboard stays good this time. []

0 Response to “It Got Around To Me”


Comments are currently closed.