Ubuntu Apache Help

[IceburgIV]

I have an Ubuntu server that I have set up that my load average are constantly over 30. I have ran through all the documents online and I can't seem to make a difference. If I run top, I see that the apache seems to be the culprit.


root@zaphod:~# ps -ef | grep -c "apache"
37


A snapshot from top shows:


17119 www-data 20 0 99.0m 13m 3448 R 7.9 0.9 0:38.61 apache2
18308 www-data 20 0 83568 60m 3024 R 7.9 4.0 0:15.62 apache2
4757 www-data 20 0 105m 52m 3452 R 5.9 3.4 14:46.99 apache2
6222 www-data 20 0 34272 14m 3480 R 5.9 1.0 8:53.42 apache2
7984 www-data 20 0 100m 17m 3564 R 5.9 1.2 9:19.96 apache2
10295 www-data 20 0 107m 34m 3484 R 5.9 2.3 9:34.79 apache2
10303 www-data 20 0 102m 21m 3620 R 5.9 1.4 5:50.44 apache2
12770 www-data 20 0 93152 31m 3708 R 5.9 2.1 6:13.88 apache2
12771 www-data 20 0 92364 29m 3568 R 5.9 1.9 7:23.33 apache2
12772 www-data 20 0 104m 31m 3524 R 5.9 2.1 5:43.97 apache2
13949 www-data 20 0 106m 32m 3460 R 5.9 2.1 5:54.75 apache2
17414 www-data 20 0 85384 14m 3284 R 5.9 0.9 0:45.32 apache2
18314 www-data 20 0 99468 76m 3020 R 5.9 5.1 0:17.32 apache2
31044 www-data 20 0 89288 22m 3564 R 5.9 1.5 18:13.54 apache2
4273 mysql 20 0 275m 71m 2788 S 3.9 4.7 246:18.88 mysqld
12778 www-data 20 0 91084 53m 3588 R 3.9 3.5 4:40.27 apache2
13948 www-data 20 0 87752 18m 3576 R 3.9 1.2 4:15.26 apache2


My apache2.conf file shows (without the commented out parts)


LockFile /var/lock/apache2/accept.lock
PidFile ${APACHE_PID_FILE}
Timeout 60
KeepAlive Off
MaxKeepAliveRequests 100
KeepAliveTimeout 4

<IfModule mpm_prefork_module>
StartServers 5
MinSpareServers 5
MaxSpareServers 10
# Was 150 SAW 20080921
MaxClients 50
MaxRequestsPerChild 0
</IfModule>

<IfModule mpm_worker_module>
StartServers 2
MaxClients 150
MinSpareThreads 25
MaxSpareThreads 75
ThreadsPerChild 25
MaxRequestsPerChild 0
</IfModule>

User ${APACHE_RUN_USER}
Group ${APACHE_RUN_GROUP}


HostNameLookups off
etc (log conf's, format's other...)


Cpu(s): 94.9%us, 0.8%sy, 1.2%ni, 2.4%id, 0.4%wa, 0.1%hi, 0.2%si, 0.0%st
Mem: 1555000k total, 845332k used, 709668k free, 3428k buffers
Swap: 3229024k total, 869096k used, 2359928k free, 142656k cached

The virtual hosts that I have are:

A Pligg site (www.neweasyrecipe.com) which only gets 100 hits / day. But the site is very intensive.

An internal picture site. Only my friends have access, and its only 1-2 hits per day.

An internal stats page. Only active on Sunday's when football is on.

An internal jinzora page. I only use it for listening to music at work.

An internal munin page for monitoring / stats of the server.

I think NewEasyRecipe has to be the issue. I can't figure out how or what to do though, or where do go. Can anyone help?

[oneloveamaru]

install "htop" it's a little bit nicer than top. Watch what process keeps popping up on top. If apache seems to be the problem, shut it down and wait a bit to see if the load goes down. /etc/init.d/apache2 stop

[IceburgIV]

I installed htop, and have attached a screen shot.

The weird thing is that if I opened another SSH window, stop apache2, the apache processes never drop out of the top or htop process list. The webpage aren't being served any more, but the top 8 processes are apache2 -k start still.

The screenshot is while apache is running. I stopped it shortly after this for about 3 minutes, before restarting it.

[IceburgIV]

No one?

[JDiss]

Without getting into technicalities, I'd probably advise you to check out your error and access logs to see if there's anything there. I did notice that htop was reporting a load average more around 16 than 30, but it's still fairly high considering what you say the traffic is like.

Are you using worker or prefork?