The Dual Pentium Pro

Well, this page is sort of an early history of my site combined with my reflections on Linux and the Pentium Pro. Notice that this page really reflects my personal opinions up until about the beginning of 2003. I've changed much about the site since this page was last updated and I'm not going to bother with documenting it here.

However, there are a couple things I must mention in the context of this page. I still dig the old Pentium Pro. It rocked then and it still rocks now, even though it's probably silly to run them to do much of anything. It was the first "real" server processor that was a cut above the usual workstation hardware. Back in the day, when you were putting together a PPro machine, it seemed as if you weren't fooling around anymore. This was big stuff and Important.

I even went as far as to try to put a quad PPro machine in service. The issue was that the old Compaq box that I scored on the cheap was very loud, didn't have redundant power supplies and although it had a nice hot-swap SCSI array built-in, it was not that fast. Did I mention it was loud?

I ended up selling it and the rest of the parts I had for it to a guy who had the space and time (and money) to see his project through.

I also ended up selling the box this page is written about. I needed the money for a project, and a guy I knew wanted the old Duallie for his home site. I had, back in 2001, decided that I was going to leave the Linux fold and start messing around with FreeBSD. The old Duallie seemed custom-made to run FreeBSD and that's what the guy wanted it for. I sold it sometime in May of 2003. It's gone from my house, but it still hosts a website on the internet, I'm told. That's good.

June, 2001 or thereabouts: OK, so I suppose somebody out there may have noticed or perhaps even cares that I'm hosting my web page from my home office. I am. It's been an uphill battle ever since I hatched the crazy idea, but it beats the free hosting sites on the internet for lack of bullshit banners and pop-ups. I might add that the learning curve was very steep.

I hatched this idea while talking to a friend of mine about how much the free web hosting site we were on sucked ass. This was largely because their business model was flawed--they kept adding customers while not upgrading their equipment. Pretty soon the site was slow, mail took forever and they experienced frequent outages. Not cool. Sure, they always had an excuse, but that really didn't matter. It's very disheartening when you're showing your page to someone and it's not there. Yes, it was free, but I got exactly the type of service I paid for.

Anyway, this has been a long journey and I've only just gone live. The big picture looks something like this:

October, 1999: Dad's dying of cancer. Life sucks. I've just finished driving for the Minneapolis River City Trolley and I don't think I'm going back. What do I do to try to get my mind off of my Dad and put my soon to be useless historical knowledge of Minneapolis History to good use? Build a website!

November, 1999: Dad's gone. I'm sad. I've written about 40 pages of historical content and I've started studying HTML. I'm also having many ideas about what else I'd like to do with the site. The site originally started as a web portal thing that would have all kinds of Minneapolis information on it. I've now realized that this was far too focused for the likes of me. I'm having difficulties trying to put the historical content together in a way that doesn't suck. v1.0waybeta is in the can by the end of the month.

December, 1999: I'm still sad, but it's nice to not have to work weekends. Holiday hustle and bustle interfere with making much progress on the site.

February, 2000: Web fever hits and I finish my HTML class in a couple of days. I've wind in my sails and I'm getting some major inspiration from somewhere. My original 8 page site balloons to about 30 pages and it keeps on growing. I go on-line officially on February 24th.

March, 2000: I start to scare up photographs I'd like to see on my site. I find a place to scan and scan the lot of them. Computer troubles make the deal difficult, but I'm off and running. I also make an icon set for the site that doesn't make me want to barf. The photos on the site are huge, but they're there.

April, 2000: Major computer problems sideline me for weeks. I keep writing, surfing and refining my code. I also start getting the first feedback from my first visitors. Major disgust follows the discovery that not one search engine has picked up my site yet.

May, 2000: Computer problems are solved. Obsession with Palm PDA begins. A newly organized web-idiot shoots 5 rolls of film in one day to fill-in holes in the site. Free hosting site troubles begin. I learn valuable lessons about crediting pictures, and reorganize the Minneapolis section.

June, 2000: The domain name tholt.com is now mine. Free site troubles continue. Redirection and a front-page hit counter tell me that people are actually hitting my site. The idea to do a daily something occurs and starts. It's getting obvious that I need to slim down my images. I see my site listed by a search engine for the first time. It's a listing of a link on someone else's site, but it's a link. A stable computer at home really helps things along.

July, 2000: I make my first .pda file for downloading. To date, I'm pretty sure nobody's downloaded any of these files from my site, but they're there for the taking. Free hosting site troubles continue. A second site is secured.

August, 2000: The image shrinking project starts and finishes. The images on my site no longer look as good as they did, but they now download fairly quickly. Site storage size goes from just over 13meg to just under 6M. I've gone over 100 pages and 100 images in the site.

September, 2000: Both free hosting sites threaten pop-up ads and one implements them. The idea for my own web server takes shape. We get DSL at home.

October, 2000: Daily column production takes it on the chin in order to free up the time it takes to learn how to install and run a Linux server. I buy my first copy of Linux. I stumble on an exceptional deal on a dual socket 8 motherboard. My first installation of Linux is finally up and stable.

November, 2000: Boxes arrive via Fedex and UPS nearly daily. I commence building a web server and a monster workstation at the same time. Web page updates become few and far between as I amass the componentry for my new boxes. A nice cash surprise towards the end of the month and I'm done buying. Monster workstation on-line.

December, 2000: Dual Pentium Pro server is up and running in its final configuration. There may be some tweaks down the line, but for now we're up and running. Switched ISP's and got a static IP. Learned how to configure my DSL router. Set up my httpd server and my named server. My wife's folks come and spend Christmas with us. I get my first hits to my host.

January, 2001: The learning curve for Linux is very steep and I've only just started feeling comfortable again with a command-line OS. The webserver is running without a hitch and the DNS server is cooking along as well. Everything else has been resisting me every step of the way. I guess that's how it goes...

December, 2001 After 12 months of trouble-free service, I rewarded this machine with a processor upgrade from 256k cache Pentium Pro chips to 512k cache Pentium Pro chips, and took it offline. Since I'm going to be re-doing the whole back end of my site, I'm going to use this box from now on as my LAN server. I'll probably install Windows 2000 Server. I need to learn it, so that's that. I find it amazing that Linux ran for 12 months without a hiccup, but that's just how reliable that OS is. The future web box will be the dual socket 7 machine I have and it will probably be running some sort of BSD.

January, 2002 Since it spent such a long time off and sitting on top of a pile of stuff, I decided to boost its RAM. It now as 384Mb of the stuff it likes. I didn't do too much else to it. It's performing very nicely.

February, 2003

It's been over a year and I recently decided that this will again be a web server. Since FreeBSD 5.0 came out, I've been dying to try it out and since this machine has been sitting, off since December, it's about time to give it a go. This machine has been sitting off for a couple of months now. I bought a bunch of stuff and built an AMD 2GHz machine for my main computer and now my PIII 800MHz is doing the job of site server. That meant that this box had to hand over the keys, so to speak. I've been waiting for some time to get rolling on this project and I was able to get going on it a couple of weekends ago.

The results were good until I tried to configure X. It crashed. Not much else to say, really. I only had a limited amount of time to mess with it that day, so I just shut it down and I'll just wait until the next batch of free time to pick it up again.

Yes, that's a long list, but you can see that these things go very quickly once they get started. I find it funny that one thing just leads to another. I get a PDA, I want to see my content on a PDA, I put my site on the PDA. It gets dull looking at it, so I decide to do a daily column. I get shitty service from a hosting site, I want to start hosting for myself. This site seems to have a momentum all of its own.

July, 2003 The Duallie box is gone. I sold it to raise money for another project that may end up being my webserver in the future. It was a nice run with this box and I'm sad to see it go.

And now, some observations:

DSL is way cool. The key to hosting from your place is some sort of permanent connection. You need a static IP to direct people to your site, and the only solution presently in this area is DSL. Surfing the internet is great at great speed and really, if you don't plan on having a billion hits, you can surf and serve your own internet site simultaneously without any performance loss. There's been a down side to getting the static IP needed to get the site off and running. Static IP's are a pain to acquire. Nobody who offers them seems particularly eager to support them and absolutely nobody I talked to from ANY ISP except the one I'm with had a clue or even an informed answer about how they were going to bill it. That is why I'm with visi.com as a provider. I called several other ISP's and these folks were the first ones to give me a straight answer.

So what am I running for hardware? As I mentioned above, It's a box I threw together in about 3 weeks after I made the decision to go forward with the project. It's a dual Pentium Pro 200/256k system that's running on Intel's Providence (PR440FX) motherboard. It's speed is pretty much exactly what a Pentium II 400Mhz chip would run. I chose this particular motherboard because it pretty much jumped off the shelf and into my lap for a very inexpensive price ($59.00 on 11/13/00). The board has tons of valuable integrated hardware and I'm making use of it all. I'm running 128M of RAM ($85.00 on 11/13/00 and $44.00 on 12/15/00) and I have a bunch of 68-pin 7200rpm SCSI drives I acquired to take advantage of the built-in SCSI on the motherboard (3x4G from Ebay for $50.00/ea on 11/26/00). I have a speedy IDE CD-ROM ($40.00 on 11/30/00) and single plain old floppy ($9.00 11/15/00).

It's in a box with a 300W power supply($64.00 11/25/00), and I popped for brand-new OEM chip fans just for reliability's sake ($15.00/each 11/13/00). It's funny that the chip fans nearly cost me more then the processors did ($24.00/ea on 10/25/00). That's not because the fans were all that expensive, it's just that the chips were so cheap. The total cost with everything except the monitor included was just over $520.00. This would have been unthinkable just a year ago because these processors were still rather expensive at that time. The bottom has fallen out of the market for Pentium Pro's with 256k cache because people are upgrading to Pentium Pro's with 1Meg of cache, Pentium Pro Overdrive chips, Pentium II Xeons, or Pentium IIIs or Athlons that would out-perform these chips for a modest outlay of cash. OK, the Xeons are still pretty pricey, but those other processors are fairly cheap for the power. I wanted a dedicated server, and Pentium Pro's (and especially dual processor setups) are server processors. Enough said.

What am I running for software? Linux, duhh. There are many good reasons for me to run Linux. First and foremost is that it was a learning experience. We have several Unix boxes at work and I really don't feel all that great about experimenting and learning about Unix on production boxes. Call me crazy, but it just wouldn't be a keen idea for me to crash our secondary email server or one of our file servers just because I kill -9'ed the wrong process. No, I needed to learn the Unix/Linux paradigm in a relatively safe environment. I really was aggressive in my quest for knowledge. I bought the esteemed book, "Linux for Dummies" and to my surprise there was a Red Hat CDROM pasted into the back of the book. I took it home, threw together a system from parts I had laying around, and was up and running in a couple days of fiddling with it. I think I had to install it about 12 times to tweak it so that it would pick up my DSL stuff and all the ancient hardware I threw at it, but it ran every time I installed it. I'd screw it up and instead of trying to fix all the stuff I'd messed up, I just reinstalled it again. I thought I might be the first person to wear out a CDROM for using it so much. It sure was a good thing I wasn't experimenting at work.

Secondly, it is possible to run Linux on less than bleeding-edge hardware and get excellent performance out of it. I can serve a webpage, surf the internet, run a GUI (graphical user interface), run an FTP server, share a printer and do all sorts of other things simultaneously all on a Pentium 133 with 32M of RAM. Try doing that with NT with any performance or with Win98 with any stability. It can't be done. Sure, I'm far more familiar with NT and Win98 and I prefer Windows 2000 to either of them, but you just can't run any of these OS'es with a single Pentium 133 processor with 32M or RAM. The only option was Linux.

The third reason was price. I wanted to serve my own website. I am not made of money. The reason I wanted to serve my own website was that I didn't want to fork out the dough to someone else to do it for me. Am I cheap? About some things, yes. I just couldn't fathom shelling out some kind of science fiction dollar-sign number to Microsoft for NT server and IIS4.0 or Win2K and IIS5. Sure, those software packages would have given me valuable knowledge that I probably could have turned into big bucks in the real world, but the company I work for doesn't host our own site and they're, frankly, less than forthcoming with the training or the money. I could probably go just about anywhere and make at least $5k/year more than I do, but I don't go. The knowledge would have been nice to wave about to new prospective employers, but I'd rather study hard, get it right and when I'm ready for massive upheaval in my personal life, find another job. I'm in the getting ready phase now.

Anyway, to bring it back to the subject, I have a few words for Microsoft: "I wanted to use your stuff, but it costs too damn much. If you aren't fearing the Linux "movement" by now, this third reason of mine is a real reason to start." I bought a full distribution of Linux, the website server Apache which 60% of the world's websites are running, and various and sundry other programs I have yet to even open, but that are very useful, all for the back-breaking price of $25.00...and it came with a book telling me how to set some of the key systems up." Wow.

The fourth reason has to do with the flexibility of Linux. Yes, I used a GUI (Graphic User Interface) to program it and set it up, but I'm really getting used to the command line stuff. The cool thing is that I can turn the GUI off when I'm not around. The fact that I can do this saves a TON of processor overhead and makes the rest of the system run more quickly. As an added bonus, there's less chance of it going boom, because the most buggy part of the system (really of any OS)--the GUI--has been turned off. Not only this, but when I need to restart a process or to adjust the settings on one of the programs running, all I need to do is make the change and then stop and restart the process. I don't have to reboot. Not having to spend time rebooting saves me downtime and that's a good thing as well. Another attractive feature of Linux is that if I don't want to run the software that comes with it, most of the alternatives are free and freely available on the internet. Choice. It's a good thing.

Hey, a guy just contacted me about linking to his Pentium Pro enthusiast site. Here's the link to PentiumPro.org. If you like the PPro as much as I do, this should be a cool site for you.