Self-inflicted slashdotting
Mar 22, 2003![[Slashdot]](/img/blosxom/technical/topicslashdot.gif){kind=small}
This morning on [Slashdot](http://www.slashdot.org), the geek forum
everyone loves to hate, there was a question regarding corporate morale.
I'm pretty happy with work now but remembering my [Management Techniques
of the Bottom 95% of U.S.
Corporations](http://www.lonsteins.com/management_techniques.html) page,
I gladly linked to it in my comment thinking,\"No one reads the 'ask
slashdot' column and my posts never get modded up.\" I was wrong.
|
Starting Score: |
1 |
point |
|
|
Moderation |
|
+4 |
|
|
50% Interesting |
|||
|
Extra ‘Interesting’ Modifier |
|
0 |
|
|
Total Score: |
|
5 |
|
Enjoy your 15 minutes. Next in line please. Oh, and be glad it was a beat column and a weekend.
I have a 1.5M/384K aDSL line and a somewhat retro collection of accumulated hardware. Without going on about all the specifics, the web server is a Sun SparcStation 5 with a model 170 mainboard. In 1997 this was pretty nice kit if not exactly a kicking machine. Expectations have changed and by contemporary standards it is pretty slow.
Expectations, however, don’t have a thing to do with doing real work. The Apache instance running on OpenBSD easily handled the load of those who read the link and those who decided to stay around and browse the rest of the site. Despite the the paltry upstream it is not taking much of a beating. It has served a little over 15,000 requests (several hundred of them CGIs) in the five hours since I posted. That works out to about 1.2 requests/second and the load average never broke 1.00. It clipped 0.83 about an hour in as log rotation kicked over the httpd process and it had to respawn all the children. Allowing the logs to grow larger before rolling and compressing took care of that. I did run ApacheBench against it last month and do some tweaking but it is unlikely that the performance is exceptional. I’m happy to inadvertantly prove the value of an old box.