12:41pm up 8 days, 22:49, 1 user, load average: 0.00, 0.00, 0.00

That's the current load average on my site, while httpd is pushing out over 5 GB of static files per day. I did see it higher before, but never above 0.20.

ps -A | grep httpd | wc shows that there are 53 httpd processes running. And here's the memory status:

total used free shared buffers cached
Mem: 257636 250564 7072 490564 9668 169744
-/+ buffers/cache: 71152 186484
Swap: 131532 1740 129792

cool. my raq using up 5 GB per month was showing server load of 20 to 80 :) all dynamic content

Hi pma

Just went to your site. You might want to check out www.ikonboard.com for FREE great forum like WHT :)

DJ: What's wrong with my current forum?

[admin admin]$ ps -A | grep httpd | wc
108 432 3240


2:19am up 14:12, 1 user, load average: 0.00, 0.04, 0.00
115 processes: 114 sleeping, 1 running, 0 zombie, 0 stopped
CPU states: 0.7% user, 3.8% system, 0.0% nice, 95.3% idle
Mem: 257636K av, 254272K used, 3364K free, 598516K shrd, 5248K buff
Swap: 131532K av, 1052K used, 130480K free 192600K cached


making 10-15 GB / day

Domus wrote:

> making 10-15 GB / day

Ack! That means you're using 300-450 GB per month. If you're using RackShack, you must be incurring a lot of bandwidth overage charges... come to think of it, if you're not using RackShack you'd also be incurring a lot of bandwidth overage charges.

How much does it cost you?

No offence, but this isn't exactly new. The Apache people say that a "low end Pentium" can saturate a 10Mbps ethernet channel; I'm not sure that I would go that far (at least, not when serving requests over the internet, which is much harder than serving requests over a LAN), but it isn't unreasonable for a single machine to be serving 50GB/day of static content.

Dynamic content, of course, is where everything slows down, and it is here that Raqs demonstrate their limits quickly.

I was somewhat surprised that the RaQ4i could handle all this load without even a noticable increase on the load average, because of all the "RaQs suck" stories that I had heard before.

I knew that a low end machine is capable of serving that much traffic using thttpd---I didn't know that it could do so using Apache too.

thttpd is faster than Apache, but it doesn't really matter very much any more. People have done benchmarks (with Zeus) showing that a single server (I think it was a 4 processor Xeon box with 16GB of RAM) can fill an OC-192 line... as long as it is only serving static content.

As for the load average... that is a really bad metric. When you approach the server's limits the load average will increase really fast. Doubling the amount of traffic can take you from a load average of 0.05 up to 5.00.

total used free
Mem: 257636 250564 7072

why is it that good, you have nearly used all your memory?!

I've looked at the status of another RaQ4i, and its memory is also all used up, even though there's hardly anything running on it.

I think it has to do with the way Linux allocates memory, and never frees it until needed. Someone else could explain this better... Basically the idea is lack of free memory isn't necessarily bad.

Originally posted by NORRITT

total used free
Mem: 257636 250564 7072

why is it that good, you have nearly used all your memory?!

A Unix/Linux system uses memory diffrently to say a windows system, If you was to put 512 RAM in side it would use this up to.

Take a look at this thread and at my post for an explanation of memory usage on linux.
http://www.webhostingtalk.com/showthread.php?s=&threadid=15129&highlight=memory+and+free

oh, ok, now I understand, thanks!

so in fact he has about 186MB free RAM! that sounds gooood :)

one is at rackshack and making around 200 GB / month
the other one is at another company offering "unlimited bandwith" well not so nice connection but at least i dont pay a dmie more i pay for raq with 256 mb ram
280 $ monthly and i make around 450 GB / month with adult movies

and no it's not kikko.com
and i will not tell the link so i could be safe that noone else would use big bandwith and then it would all suck or they would change the policy
they are ISP like EV1.net and they have 20 mbit line to CW.net
so plenty of space for me :) and few friends

> the other one is at another company offering "unlimited bandwith"
> and i will not tell the link so i could be safe that noone
> else would use big bandwith and then it would all suck
> or they would change the policy

Hey, do you guys think it's my fault that ServerCentral.net's unlimited bandwidth sucks now? (I posted about them here and on GeekTalk)

A few weeks ago, they installed a new kernel upgrade into their FreeBSD virtual servers that make thttpd not work properly anymore, and also made the CPU time limit stricter. This artificially limits the amount of bandwidth we can put out.

got systems warnings like this daily :/

--
& Message 4:
From root Wed Aug 22 14:45:15 2001
Date: Wed, 22 Aug 2001 14:45:10 -0700
From: admin
To: sys
Subject: The CPU is heavily loaded

Over the past fifteen minutes, the CPU has been heavily loaded.

This will result in noticible performace loss. Consider moving some of the
services to other Cobalt servers, or reduce the complexity of the CGI
scripts running on the Cobalt server itself.

1 minute load average: 17.13
5 minute load average: 11.84
15 minute load average: 7.15
--