Hi,

think about replacing apache with lighty (www.lighttpd.net)
ist faster, more stable and does not need as much memory as apache.

(only ONE) Example:
Apache2 website has 200+ req/sec 1200ppl logged in concurrent load 200+
SAME machine with lighty and same conditions (requests and logged in ppl)
load UNDER 2.

It’s the better choice. ;)

Simon


-----Ursprüngliche Nachricht-----
Von: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Im Auftrag von x-ares-x
Gesendet: Freitag, 13. Januar 2006 17:34
An: hlds_linux@list.valvesoftware.com
Betreff: Re: [hlds_linux] Load Average

Hey Luca,

Thanks for the mail, I shut the apache down for a test.
But I have to start it again i 2 hours.

I'm running HLstatsX, so there may be a lot of mySQL queries.
Maybe I should test it without the stats.

It's really not unplayable, bit there are lags from 1 second, sometimes.
when the load rises about 3-4.

Don't know, if the configuration will help or more performant hardware,
like a new CPU or additional RAM.

And about the too many processes:
I'm new with Linux, so I don't know every process that can be disabled.
But Apache got 10 Child's and Teamspeak too.
Here some Tasks in the moment mysql was active:

2795 root      15   0  241m 138m 1124 S 20.0 31.4 566:24.58 srcds_amd
13229 root      15   0  210m 115m  936 S  7.8 26.2 258:50.05 srcds_amd
  1753 mysql     16   0  122m 6216 2936 S  1.7  1.4  12:05.62 mysqld
15824 root      16   0 13732 9312 2632 S  0.9  2.1   0:11.64 hlstats.pl
  3773 root      16   0  2012  980  780 R  0.3  0.2   0:00.05 top
25100 css_admi  16   0  7428 1632 1148 R  0.1  0.4   0:01.27 sshd
  1519 root      16   0  2740  456  388 S  0.0  0.1   0:49.28 nifd
  6185 css_admi  15   0 23244 1584 1112 S  0.0  0.4   1:50.12 server_linux
  6187 css_admi  16   0 23244 1584 1112 S  0.0  0.4   2:57.50 server_linux
     1 root      16   0  1740  488  460 S  0.0  0.1   0:02.45 init
     2 root      34  19     0    0    0 S  0.0  0.0   0:00.00 ksoftirqd/0
     3 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 events/0
     4 root      19  -5     0    0    0 S  0.0  0.0   0:00.00 khelper
     5 root      11  -5     0    0    0 S  0.0  0.0   0:00.00 kthread
     7 root      20  -5     0    0    0 S  0.0  0.0   0:00.00 kacpid
    71 root      10  -5     0    0    0 S  0.0  0.0   0:00.15 kblockd/0
    74 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 khubd
   121 root      15   0     0    0    0 S  0.0  0.0   0:00.85 pdflush
   122 root      15   0     0    0    0 S  0.0  0.0   0:03.69 pdflush
   124 root      16  -5     0    0    0 S  0.0  0.0   0:00.00 aio/0
   123 root      15   0     0    0    0 S  0.0  0.0   0:02.18 kswapd0
   211 root      10  -5     0    0    0 S  0.0  0.0   0:00.05 kseriod
   328 root      11  -5     0    0    0 S  0.0  0.0   0:00.00 kmirrord
   338 root      15   0     0    0    0 S  0.0  0.0   0:53.11 kjournald
   763 root      12  -4  1636  456  452 S  0.0  0.1   0:00.06 udevd
   795 root      19   0     0    0    0 S  0.0  0.0   0:00.00 shpchpd_event
   972 root      10  -5     0    0    0 S  0.0  0.0   0:02.20 kauditd
  1103 root      15   0     0    0    0 S  0.0  0.0   0:00.00 kjournald
  1415 root      16   0  1612  516  468 S  0.0  0.1   0:29.55 syslogd
  1417 root      16   0  1568  316  312 S  0.0  0.1   0:00.01 klogd
  1425 rpc       15   0  1692  408  404 S  0.0  0.1   0:00.00 portmap

Luca Fabbro schrieb:

> At 16.41 13/01/2006, you wrote:
>
>> Sure There are other processes,
>> but they all have no CPU usage,  0.1 Sometimes.
>> I don't know how to Find the process that prodices the Load...
>>
>> Do you want my Processtable?
>
>
> No, I won't be able anyway to tell you why load average goes up if I
> don't have a shell (and I'm not sure that I'll be able even if I have one)
>
> Having a quick look at the graphs load average of 1 doesen't seem so
> strange.
> In fact you have a raw cpu usage in user space of around 38 and
> system is around 6
> This brings the typical usage to 44% of your CPU. I don't know if
> this is all generated by the servers.
> But you have a high number of processes active (average 91). As the
> load average takes into account all the processes when apache for
> example spawns or uses one of it's child processes (and it seems that
> you have some requests on your web server) the computation of the
> load average will grow, and it will be more than one.
> This doesn't mean that the server will be unplayable but it's a big
> warn regarding the ability of the machine to run all the "possible"
> tasks in a smoot way.
> Removing one of the server processes will reduce the basic cpu usage
> and also the average load will be for sure reduced.
> I think that if you stop the web server (but I don't think you can)
> you'll se the load average back below 1
>
> Hope that this helps
>
>
>
>> Luca Fabbro schrieb:
>>
>>> At 15.38 13/01/2006, you wrote:
>>>
>>>> Hey Fox, and the others,
>>>>
>>>> yes I know that when loadaverages is at 1, the server is running at its
>>>> maximum. But I got peaks with 3-4 ...
>>>
>>>
>>>
>>> Load average can be more than one. Load average tries to tell you the
>>> probable "ability" of the processes of you machine to ran smoothly
>>> (that's a lot simplified).
>>> On your machine there are also other processes as for example apache
>>> ones.
>>>
>>> googling
>>> http://www.teamquest.com/resources/gunther/ldavg1.shtml
>>>
>>> Here you can find an explanation to your question.
>>>
>>>> So the Server only got 1 CPU, its an AMD 2200, the Problem is, thats it
>>>> is a Sempron.
>>>> I've got a 2.6 Kernel with 1000Hz rate
>>>> And only 512MB RAM, but Linux was always using 260MB. Now with a 3rd
>>>> server up to 320MB.
>>>> The System is not very great, but I don't had to pay for it.
>>>>
>>>> Here you can look at some graphs: http://www.o711.com/stats/
>>>> The traffic last night was made by 8 playing people.
>>>>
>>>> For now I will stop the 3rd Server, an look the effect.
>>>>
>>>> I think I don't really understand what is going on on my System.
>>>>
>>>> Ohh, I forgot.. its a 34Mbit Connection.
>>>>
>>>> Sorry for my bad english... :-)
>>>>
>>>>
>>>> Reinder P. Gerritsen schrieb:
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> To unsubscribe, edit your list preferences, or view the list archives,
>>> please visit:
>>> http://list.valvesoftware.com/mailman/listinfo/hlds_linux
>>>
>>>
>>
>>
>> _______________________________________________
>> To unsubscribe, edit your list preferences, or view the list
>> archives, please visit:
>> http://list.valvesoftware.com/mailman/listinfo/hlds_linux
>
>
>
> _______________________________________________
> To unsubscribe, edit your list preferences, or view the list archives,
> please visit:
> http://list.valvesoftware.com/mailman/listinfo/hlds_linux
>
>
>


_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux


_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux

Reply via email to