I have an Ubuntu server machine running Apache and MYSQL. System and version info is as follows: Linux kernel 3.0.0.-12 Apache/2.2.20 MySQL Ver 14.14.Distrib 5.1.58
I am running a few websites on this server, some HTML only, some PHP/MySQL. THe [problem is that response time is very slow, both on static as well as the dynamic sites. Sometimes it takes more than 10 seconds before a response is given, this makes the sites very slow and almost unusable.
The problem occurs even when requesting from the local network. I have added the involved subdomains to my /etc/hosts file, and abolve all the problem is not solved by using IP numbers instead of URL's. So there is no DNS lookup issue.
I have modified the log format by showing the response times and sometimes a files takes 12 seconds to be served, see the jquery~.js file in the example screenshot.
I have no explanation for this extremely long response time, but is is not even the only issue here, some other files takes a long time to be served too, but do not show a long response time in the log file.
So probably different tissues are involved here. I cannot find a solution until now, any suggestions???
THanx in advance, Klaas
link to screenshot picture from access logfile
Some extra configuration info:
apache2.conf (comment is removed)
LockFile ${APACHE_LOCK_DIR}/accept.lock
PidFile ${APACHE_PID_FILE}
Timeout 300
KeepAlive On
MaxKeepAliveRequests 100
KeepAliveTimeout 5
<IfModule mpm_prefork_module>
StartServers 5
MinSpareServers 5
MaxSpareServers 10
MaxClients 150
MaxRequestsPerChild 0
</IfModule>
<IfModule mpm_worker_module>
StartServers 2
MinSpareThreads 25
MaxSpareThreads 75
ThreadLimit 64
ThreadsPerChild 25
MaxClients 150
MaxRequestsPerChild 0
</IfModule>
<IfModule mpm_event_module>
StartServers 2
MinSpareThreads 25
MaxSpareThreads 75
ThreadLimit 64
ThreadsPerChild 25
MaxClients 150
MaxRequestsPerChild 0
</IfModule>
User ${APACHE_RUN_USER}
Group ${APACHE_RUN_GROUP}
AccessFileName .htaccess
<Files ~ "^\.ht">
Order allow,deny
Deny from all
Satisfy all
</Files>
DefaultType text/plain
HostnameLookups Off
ErrorLog ${APACHE_LOG_DIR}/error.log
LogLevel warn
Include mods-enabled/*.load
Include mods-enabled/*.conf
Include httpd.conf
Include ports.conf
LogFormat "%v:%p %h %l %u %t \"%r\" %>s %O \"%{Referer}i\" \"%{User-Agent}i\"" vhost_combined
LogFormat "%h %l %u %t \"%r\" %>s %O \"%{Referer}i\" \"%{User-Agent}i\" %T/%D" combined
LogFormat "%h %l %u %t \"%r\" %>s %O" common
LogFormat "%{Referer}i -> %U" referer
LogFormat "%{User-agent}i" agent
Include conf.d/
Include sites-enabled/
And the virtual hostfile for one of the slow sites, in fact it is pretty straightforward...
<VirtualHost *:80>
ServerAdmin [email protected]
ServerSignature EMail
ServerName toenjoy.drsklaus.nl
DocumentRoot /var/www/toenjoy.drsklaus.nl
<Directory />
Options FollowSymLinks
AllowOverride None
</Directory>
<Directory /var/www/toenjoy.drsklaus.nl/>
Options Indexes FollowSymLinks MultiViews
AllowOverride AuthConfig
AuthType Basic
AuthName "To Enjoy"
AuthUserFile /etc/.htpasswd
Require user petraaa
Order allow,deny
allow from all
</Directory>
ScriptAlias /cgi-bin/ /usr/lib/cgi-bin/
<Directory "/usr/lib/cgi-bin">
AllowOverride None
Options +ExecCGI -MultiViews +SymLinksIfOwnerMatch
Order allow,deny
Allow from all
</Directory>
ErrorLog /var/log/apache2/error.log
# Possible values include: debug, info, notice, warn, error, crit,
# alert, emerg.
LogLevel warn
CustomLog /var/log/apache2/access.log combined
Alias /doc/ "/usr/share/doc/"
<Directory "/usr/share/doc/">
Options Indexes MultiViews FollowSymLinks
AllowOverride None
Order deny,allow
Deny from all
Allow from 127.0.0.0/255.0.0.0 ::1/128
</Directory>
</VirtualHost>
And the output of free -m:
klaas@ubuntu-server:/etc/apache2$ free -m
total used free shared buffers cached
Mem: 1997 1401 595 0 144 1017
-/+ buffers/cache: 238 1758
Swap: 2035 0 2035
and I have no indication that swapping occurs on the moments the site is slow. I have runned top and it does not appear to be a CPU issue. I have the impression that the spawning of a apache thread could maybe be the bottleneck but it is just a suggestion. Maybe this gives some extra information!
EDIT: The problem seemed to be gone for some time but occurs again! And not only with Apache, also connecting using SSH takes a tremendous time, sometimes it takes up to 15 seconds before the keyphrase is asked for. Also scp works very slowly. The behavious is really unpredoctable and makes the server very hard to use. Any ideas...?
Edit: Try these one at a time!
A couple of things jump out:
If you have no reason to use
0
(process never dies) then increase it by a little (default is1000
) to say100
and restart, and see if that gets you improvements. May prevent memory leaks.Depending on how much memory your server has, and if all three modules are being called at at the same time, you may run out of memory. For e.g., if your apache process is 20MB in size, you will have 150 Clients x 20MB x 3 mods = 9000 MB ~ 8.8 GB. Even if only one of them is being called, then you are using up ~ 2.9 GB.
If you are not using all three mods, uninstall the unneeded ones from your apache instance.
Also, how about the output oftop
andfree -m
during this slow performance?What happens when you attempt to access via
lynx
orcurl
from the local host or from a machine on the network?Also, look at this block from Apache:
I would recommend getting rid of that whole block, restarting Apache & trying again. I have had odd hangs as well when Apache does some
deny
actions against a localhost or a local IP address. Weird DNS lookup issues can hang Apache on a per request basis. And a delay of 12 seconds per page makes me think that is the cause. But it's not 100% clear from your config if the block I am pointing out is the culprit or not. But do some kind of search for DNS lookup related tasks connected to Apache.Also, it's hard to say what settings should be on your server without observing your sites real traffic, but I suggest that your
KeepAlive
settings seem resource heavy. I would recommend changing as so:I would also recommend lowering
MaxClients
and addingServerLimit
adjustingMaxRequestsPerChild
as well in your mainapache2.conf
like so:Also, I highly recommend installing Munin on your machine to get a grip on resource usage. Saves a lot of headaches & gives a good broad perspective on things.
Apache is a complex beast and may be the problem here...or may not. I would start debugging the general network slowness before reaching into that dark hole. Specifically:
Try running netcat in server mode on a port, configured to hand out a local file's contents, and then connecting to it from the machine itself. For example:
while true ; do nc -l 80 < index.html ; done
That eliminates apache, the physical network, and DNS or IP link issues. Note that the file served is local - no NFS mounts for example.
If you get this far, your server is good, and your network routing is OK too.
Now clean up DNS, and make quite sure that reverse lookups work. dig 4.3.2.1.in-addr.arpa for example to test the lookup of the 1.2.3.4 address. If it's slow, that's a problem and you'll need to noodle about with /etc/resolv.conf, dnsmasq or whatever you have. Don't proceed until this is rock solid. Worth trying SSH/SCP into the box at this point too - it should be fast.
Next, Apache. If you make it this far you know your network is good, DNS works and the box is OK. So that leaves your application (ie the www server). Try to strip down the config to the bare minimum and see how you get on. For example here is one.
Good luck!