Computers aren’t stupid

Sometimes I have a feeling that computers aren’t that stupid. They know things. Here is a fresh example.

My office workstation started misbehaving. During the last few days it got to alsmost unbearable. It hangs every 20 minutes or so. I did all the troubleshooting and debugging and everything looks OK. I’ve monitored the motherboard and CPU temperatures – neither ever got above +34C. I ran a bunch of tests on the memory – everything is good. I checked both harddisks for bad blocks – none found. I removed all unused hardware and drivers. I have changed drivers for my video card, which is a normal NVidia GeForce MX 32. I did nothing fancy with the PC.

Still, about every 20-30 minutes it would hang totally dead. I got bored with the situation and ordered a replacement. One was going to happen sooner or later anyway, since my current workstation is pretty old.

We all know what jealosy is, don’t we? Well, guess what happened. The very next day after my purchase order was approved, I come to the office, login, and my workstation crashes on me. I reboot it as usual, but it refuses to come up. I get a kernel oops. And a one I haven’t ever seen before. Yey! I boot with the rescue CD and realize that my root filesystem is terribly broken. Coincidence? I think not.

fsck.ext3 was running for more than two hours trying to repair everything. But that didn’t help. The machine is coming up to some really strange state – it does not load any services on the startup an proceeds directly to the login screen. It does not allow in noone, even local root. Plenty of errors are getting dumped on the console about missing shared libraries and stuff like that.

In about one hour I’ll be out of ideas…

P.S.: I’ve mentioned the Murphy’s Law recently. Well, it struck again. Our backup library died a few days ago and we are in the process of replacing it. So even if there is any fresh backup of my computer, I can’t restore it at the moment. Cool!

6 thoughts on “Computers aren’t stupid”

  1. From reading past articles I think you are using Red Hat or Fedora. If you can’t log in as root then either /etc partition (which I doubt you made a separate partition for) or /usr is messed up. /usr/lib is where PAM libraries are for authentication. However you should still be able to get into root in runlevel 1 (if not there is always rescue). runlevel 1 bypasses PAM authentication.

    At that point you should be able to backup the data but it sounds like you may need to do a fresh install then a restore.


  2. Psuche,

    Yes, you are right – I am using Fedora Linux. My /usr and /etc are on the root partition and that one got messed up pretty bad. The machine is unbootable (none of the runlevels). I’ve attempted to revive it with the rescue, but it is easier to reinstall it.

    I can access all of my critical files on /etc and /home. Everything else I don’t much care about.

    The question now is how long should I wait. There are two issues to consider:

    1. My new workstation is on the way. It might get here anywhere within a week. I am thinking about using a terminal until then and not attempting to reinstall the current machine.

    2. Fedora Linux Core 4 is about to be released within the next few days. I will surely use it. What I want to avoid is Core 3 installation and than upgrade to Core 4 on the next day.

    Anyway, I’ll see how things are on Monday. In the worst case scenario, using a terminal is not such a bad thing for a few days anyway. :)

  3. In that case download a knoppix iso or something that has the apps you want. If you can get the data you need later then just wait.

    Laziness is the mother of invention ;)

Leave a Comment