"Linux Gazette...making Linux just a little more fun!"

Horror Story

Here's a quick horror story for you. I was recently admin-ing one of my Linux servers. This server is the company print server. What had happened was a user was connected to a legacy system via a terminal program. Somewhere along the line there was a screwup and his system starting dumping lots of extraneous data to the nearest printer. The user tried to power-cycle the printer to no avail. Thankfully, they came to me next to try and resolve the problem.

Anyway, I ssh'd into the Linux box, changed to the 'spool/' directory and did an 'ls' of the appropriate folder. After noting that there were a few hundred print jobs waiting in the print queue I decided it was best if I just deleted them all...

Here is where the proverbial hit the spinning thing!

...I decided to 'rm -rf *' all the files. The problem is I did it from the 'spool/' directory rather than the proper printer directory. This had the effect of deleting everything in my system spool directory!

After realising my error I tried to fix it. First thing was trying "unrm". This did not work as I could not read from the filesystem without errors. Next I tried using a boot floppy and trying 'unrm' again. But no that would not work because I could not get the RAID array to be recognised!!! Tried a few other options until I gave up and decided it was time for an "upgrade".

Problem solved after 3 hours :)

After buying an all-on-board style PC and installing linux, I wanted to try out a dual-headed configuration. This was mainly because I found myself with a spare graphics card and monitor. I plugged in the new graphics card, and realised that the super complex modular BIOS system was automatically disabling interrupts from the on-board graphics system. According to the manual, it was impossible to turn off this 'feature'. Then I had an idea!

The idea was that if the new graphics card could be disabled while booting, the BIOS may ignore it. Then, I could re-enable the card for normal use.

After studying the PCI bus pinout and specs, I decided to disconnect the reset line from the graphics card using a sharp screwdriver to destroy the track. I then re-connected the reset line through a switch, and then down to the ground connection. (Reset is Active Low)

Using the switch I could disable the card, but it then became impossible to re-enable it because the PC could not reset it when it wanted to. I then decided to connect a resistor across the broken track so that the card could also be reset by the PC, and my switch would not affect anything else.

The switch was neatly mounted on the metal back plate of the card, and can be switched when required.

Using the switch, I put the card into reset mode, booted the PC, then turned the card back on, and found myself with two working monitors ;-) ;-)

Also, have you ever wondered how to get out of the situation where X does not return VGA text mode? Adding one of these switches to your card solves this problem.

[If you have a story about something foolish or ingenious you did to your computer, send it to [email protected]-Iron.]

Mike Orr

Mike ("Iron") is the Editor of Linux Gazette. You can read what he has to say in the Back Page column in this issue. He has been a Linux enthusiast since 1991 and a Debian user since 1995. He is SSC's web technical coordinator, which means he gets to write a lot of Python scripts. Non-computer interests include Ska/Oi! music and the international language Esperanto. The nickname Iron was given to him in college--short for Iron Orr, hahaha.

---

Copyright © 2002, Mike "Iron" Orr.
Copying license http://www.linuxgazette.net/copying.html
Published in Issue 76 of Linux Gazette, March 2002

---