« Patrick's Rules? | Main | Does your weblog own you? »

September 09, 2003

Unix Geek - History Continues

It was around this time that I used to play a lot with PC hardware. True, I mainly used it to run various forms of Unix but it was still PC hardware. The 386 was coming out, DR-DOS was all the rage especially if you added memory management software. People were arguing about which way to manage extra memory - should you use Expanded Memory or Extended Memory. There was all these fun things like DMA and interrupts to worry about and "smart" cards to compensate for inadequate hardware (smart disk cards with lots of RAM, smart serial cards that supported 422, smart video cards that could get more than 640x480). The best bit about the whole process was that nothing was supposed to work together so when it did you felt really great. The downside to the "plug and play" revolution was that things were supposed to work (in reality the probability that it would had not changed) and so you feel cheated when it doesn't. The whole thing comes down to an expectation. I used to have a few friends in the PC assembly business and I was forever trading or testing gear to get the "ultimate" PC. I played a few games but the real game was getting things to work.

Now things weren't so good at the lab. There were layoffs and reorganisations. The migration project was delayed (again) and the cobble-together solutions were getting less interesting. I started to look for a new opportunity and eventually found a really cool job. There used to be a businessman who would travel to the US a couple of times a year with some of his staff to select possible products for the Australian market. He would pick a company/product, set up an Australian subsidiary with a related name, build the business and then sell it back to the parent at a great profit. He had a real nose for the best IT products and had previously created Wicat Australia, Prime Australia and more recently Sun Australia. He had just sold Sun Australia back to Sun Microsystems and was now investing in (amoung other things) Pyramid (high end Unix), and NCD Xterminals. There was a small problem that some former customers still had support agreements with his company that had not transferred to Sun but all the Sun staff had of course been part of the sale to Sun. I came on board to support the remaining Sun customers and to help with the Xterminal and X-windows software markets. Some parts of the job were cool - I got to play with the brand new Sun Sparc boxes (the Sun 4) as well as lots of older hardware (Sun 3/60, Sun 3/80, Sun 4/110) and really cool X software. Other parts of the job were not so good, I had a new boss and a new business card every three months or so and there was almost too much to learn.

I was technically a pre-sales engineer with some support work and I have some great customer stories about this period. There was the customer who decided to make Unix more friendly and so he renamed things so that /bin became /commands, /usr became /user and /etc became /other-stuff. Problem was this SunOS 4.0.3 system refused to boot after the changes. Funny that. Another customer was panicing because of a shortage of space in the root filesystem. After checking, (s)he identified /vmunix as the largest file and compressed it. Once again, no boot! I guess non-unix people won't appreciate these but I can still remember the pain of fixing them. The OS was on tape, usually two of them. The first one created a mini-unix in memory and the second (and later third) did the install. You could boot off the first tape and from the mini-unix mount and repair such errors BUT if you wanted to add your favourite diagnostic tools, you would have to cut a new boot tape which was (a) complex and (b) error-prone. Once an engineer had a good boot tape with utilities, he would never let it go (I think I still have mine somewhere) :-)

Another customer had this tendency to ring up and ask strange questions without really providing any context. This invariably led to trouble, as an example there was a call about handling arguments in a shell script. The next thing I knew was a call from my manager about bad advice.... turns out the customer was writing a 'del' script to replace the Unix unfriendly 'rm' command. The shell script was parsing the arguments but instead of using "$*" as recommended, was using "*" and in fact had ended up being effectively: rm -rf * which forecfully and recursively removes everything. Customer had test this script while logged in as an administrator in the root directory. Fortunately, my co-workers (or is that cow-orkers) had explained this to the boss by the time I got back to the office (a few days later).

I tried to institute an award system, the "Richard Cranium" award for customers but after a bit of fanfare, someone worked out what it really meant. Maybe it was the list of possible awardees posted on the notice board. Anyway, management decided that it was not a very "customer focused" attitude. It was around this time that I came across the BOFH and he has been my role-model ever since :-)

That is all the time I have for this article because (checks BOFH excuse):

Communist revolutionaries taking over the server room and demanding all the computers in the building or they shoot the sysadmin. Poor misguided fools.

To be continued ...

Posted by Ozguru at September 9, 2003 08:09 AM


Comments


I have to admit that I did not invent the "Richard Cranium" award. It was awarded at univeristy to a student who printed the fortune program on the band printer. This was a "very bad thing"(tm) because the file had a binary section (the hash table) followed by some text. The binary section used to driver the printer nuts and in this case had used an entire box of paper (they were big boxes full of continuous 132 column stationary). The student had then gone and got another two boxes to feed the printer - even though the outpuit was pure gibberish. The award was given at an end-of-year function and the fellow got all the way up to the front before he managed to work out who the distinguished Dr Richard Cranium PhD really was :-)

Posted by: ozguru at September 9, 2003 08:09 AM