Skip to main content

performance - Apache consuming too much CPU and memory



I am having some troubles with CPU loading an memory with Apache Web Server.



We are running a Ubuntu Server 12.04 LTS on a Virtual Machine. Our server have the following specs:




  • 8GB RAM;

  • 4 vCPUs (12ghz);




We configured the server to run a Drupal (7.23) based website. So, we installed Apache, PHP, MySQL... The versions are below:




  • Apache 2.2.22;

  • PHP 5.3.10 (The PHP are running as Apache Module.);

  • APC 3.1.7;

  • MySQL 5.5.31 (all innodb tables);




I am running some apache modules too. Take a look (apachectl -M):




  • core_module (static)

  • log_config_module (static)

  • logio_module (static)

  • mpm_prefork_module (static)

  • http_module (static)

  • so_module (static)


  • actions_module (shared)

  • alias_module (shared)

  • authz_host_module (shared)

  • deflate_module (shared)

  • dir_module (shared)

  • env_module (shared)

  • include_module (shared)

  • mime_module (shared)

  • php5_module (shared)

  • proxy_module (shared)


  • proxy_http_module (shared)

  • reqtimeout_module (shared)

  • rewrite_module (shared)

  • setenvif_module (shared)

  • ssl_module (shared)

  • status_module (shared)



On apache2.conf, we have this config:




    Timeout 90
KeepAlive On
MaxKeepAliveRequests 80
KeepAliveTimeout 5
HostnameLookups Off
LogLevel warn


StartServers 10
MinSpareServers 10

MaxSpareServers 30
MaxClients 120
MaxRequestsPerChild 1000



The Virtual Host of my site:



    
ServerName blabla.bla.bla

ServerAdmin bla@bla.com
DocumentRoot /l/disk0/site/public_html


AllowOverride None



Options MultiViews Indexes Includes FollowSymLinks ExecCGI
AllowOverride All

Order allow,deny
allow from all


LogLevel warn
ErrorLog "/l/disk0/site/logs/apache/site/error.log"
CustomLog "/l/disk0/site/logs/apache/sit/access.log" combined
SSLProxyEngine on
RewriteEngine on
RewriteLog logs/rewrite_www_log

RewriteLogLevel 0

Include rewrites-www.conf



Drupal Modules:




  • ACL 7.x-1.0


  • APC - Alternative PHP Cache 7.x-1.0-beta4

  • Boost 7.x-1.0-beta2

  • Cache Expiration 7.x-2.0-beta2

  • CAPTCHA 7.x-1.0

  • Chaos tool suite (ctools) 7.x-1.3

  • Date 7.x-2.6

  • Domain Access 7.x-3.10

  • Domain Blocks 7.x-2.0

  • Domain CTools 7.x-1.3

  • Domain Locale 7.x-1.0-beta3


  • Domain Taxonomy 7.x-3.x-dev (2012-abr-29)

  • Domain Views 7.x-1.5

  • Embed Views Display 7.x-1.2

  • Entity API 7.x-1.2

  • Entity reference 7.x-1.0

  • IMCE 7.x-1.7

  • IMCE Mkdir 7.x-1.0

  • Internationalization 7.x-1.10

  • Link 7.x-1.1

  • Localization update 7.x-1.0-beta3


  • Media 7.x-1.3

  • Meta tags quick 7.x-2.7

  • Newsletter 7.x-1.0-beta9

  • Options Element 7.x-1.9

  • Page Style 7.x-1.0

  • Panels 7.x-3.3

  • Pathauto 7.x-1.2

  • pathologic 7.x-2.11

  • profile2 7.x-1.3+0-dev (2013-mai-24)

  • select_or_other 7.x-2.19


  • sheetnode 7.x-1.0-beta4+3-dev (2013-mai-25)

  • SMTP Authentication Support 7.x-1.0

  • Token 7.x-1.5

  • Transliteration 7.x-3.1

  • Variable 7.x-2.3

  • Views 7.x-3.7

  • Vocabulary Permissions Per Role 7.x-1.0

  • Webform 7.x-3.19

  • Webform Validation 7.x-1.2

  • workbench 7.x-1.2


  • workbench_access 7.x-1.2

  • workbench_media 7.x-1.1

  • workbench_profile 7.x-1.1

  • xmlsitemap 7.x-2.0-rc2



My site is simple and don't have many visitors. I am talking about 500 visitors a day maybe. Drupal can cause so much CPU loading? Or a module?



Other problem is memory usage. When a process is created, 80M is allocated for apache2. I think is too much.




My problem is that CPU (all cores) have a high load. Most of time, hitting between 90% and 100% load! The offending process is the apache2. Memory is also consumed without pity. Of a 8GB total, the consumed memory is about 6.5GB to 7.5GB. I don't know if my apache configuration is wrong or if I'm really need more hardware (I guess not). Drupal can cause high CPU load?



When the CPU load hits 100%, the site goes down and we have to restart apache. I did a workaround solution with Drupal using APC and installing Boost. had some effectiveness, but CPU load still high. Very high.



If you need more information, like Drupal modules and PHP extensions. Please let me know.


Answer




Other problem is memory usage. When a process is created, 80M is allocated for apache2. I think is too much.





Is that real or virtual memory? Honestly, it's not very much; more to the point, you should focus on fixing things that are causing problems, not just things that you "think" should be different.



If you want Apache processes to take up less memory, you should disable modules (as each one is more code that needs to be in memory). But if you need all the modules you've got enabled, then, well, that's that.



One approach I've used when administrating a memory-constrained machine was to move certain tasks out of Apache and into other servers, so I could tune them separately.



But a far simpler approach is to change



MaxClients           120



to something more reasonable for your workload:




If we consider the rewrite accesses and the main site traffic, we have about 70 requests per minute. Right now, we have 33 incoming connections.




I'll get back to that momentarily, but if you're only dealing with 33 concurrent requests, you don't need 120 workers!



MaxClients           40



And you should probably tune down MinSpareServers and MaxSpareServer to something like 5 and 10, respectively. There's no need to have 30 workers sitting around doing nothing.



Now, getting back to




If we consider the rewrite accesses and the main site traffic, we have about 70 requests per minute. Right now, we have 33 incoming connections.





If you've got 33 concurrent requests, but you're only doing 70 a minute, there are a couple of possibilities:




  1. Your requests are taking around 30 seconds each to serve!

  2. Your request rate isn't very stable, and most of the minute you're doing nothing.



If #1 is the case, I don't really know how to help - something is incredibly wrong, so wrong I don't even know where to tell you to start looking.



If it's #2, my guess is you're serving all your static assets (images, js, css, fonts) from your server. It's best to put these on a CDN, but if you really can't do that, you can set super-long cache times on them and turn Varnish back on. If you're using Apache processes with PHP and a whole host of other things just to serve static files, you're wasting resources - do that with something simpler!





My problem is that CPU (all cores) have a high load. Most of time, hitting between 90% and 100% load! The offending process is the apache2.




Is this a constant number, or only when you're serving requests?



How does disk I/O look (iostat -mhx 2)? What is MySQL doing (show processlist;)?







Your server is vastly overpowered for what you've described. This is good news, because it means you should be able to fix this problem.


Comments

Popular posts from this blog

iLO 3 Firmware Update (HP Proliant DL380 G7)

The iLO web interface allows me to upload a .bin file ( Obtain the firmware image (.bin) file from the Online ROM Flash Component for HP Integrated Lights-Out. ) The iLO web interface redirects me to a page in the HP support website ( http://www.hp.com/go/iLO ) where I am supposed to find this .bin firmware, but no luck for me. The support website is a mess and very slow, badly categorized and generally unusable. Where can I find this .bin file? The only related link I am able to find asks me about my server operating system (what does this have to do with the iLO?!) and lets me download an .iso with no .bin file And also a related question: what is the latest iLO 3 version? (for Proliant DL380 G7, not sure if the iLO is tied to the server model)

linux - Awstats - outputting stats for merged Access_logs only producing stats for one server's log

I've been attempting this for two weeks and I've accessed countless number of sites on this issue and it seems there is something I'm not getting here and I'm at a lost. I manged to figure out how to merge logs from two servers together. (Taking care to only merge the matching domains together) The logs from the first server span from 15 Dec 2012 to 8 April 2014 The logs from the second server span from 2 Mar 2014 to 9 April 2014 I was able to successfully merge them using the logresolvemerge.pl script simply enermerating each log and > out_putting_it_to_file Looking at the two logs from each server the format seems exactly the same. The problem I'm having is producing the stats page for the logs. The command I've boiled it down to is /usr/share/awstats/tools/awstats_buildstaticpages.pl -configdir=/home/User/Documents/conf/ -config=example.com awstatsprog=/usr/share/awstats/wwwroot/cgi-bin/awstats.pl dir=/home/User/Documents/parced -month=all -year=all...

linux - How can I get my mediawiki to stop thinking I have cookies disabled?

I've searched half a day for how to resolve this issue, and can't figure it out. Shortly after I made my wiki a simple private wiki according to the instructions at Mediawiki's website, it started giving me this weird login error message: Wiki uses cookies to log in users. You have cookies disabled. Please enable them and try again. If I remove those private wiki settings, the error disappears, even if I try logging in. But I need it to be a private wiki for only my team. So what do I do? Here's what I've done so far. Just to be safe, after ever change, I try rebooting Apache using: sudo /etc/init.d/apache2 restart In my php.ini file, I have the following set: session.save_path = "/var/lib/php5" session.cookie_secure = secure session.cookie_path = /tmp session.cookie_domain = my server's internal URL (should I even set this? this field was blank before, but not commented out) session.referer_check = Off I ran the following to ensure that the fold...