By albertc on
Hi
Since I moved my Drupal site to production, my dedicated server is experiencing repeated failures in httpd, mysqld and cpsrvd, so not only the site itself but everything else in the server becomes unavailable.
The hosting people tell me this:
During the time of apache / cpsrvd restart , the server uses above 95% of the memory with mysql & apache being the most resource eating processes. Also during this period ,the mysql user MYDBUSER is having many queries in 'Query' state which means the database server is busy.
Can somebody suggest a solution?
Thank you in advance for any assistance
--
Albert
Comments
You may want to explain more
You may want to explain more about what you have running on your server. Is it just one website with drupal? What modules are you running? Who many visitors are you receiving at any moment/period of time? How much content is on your drupal installation? There are so many factors that can come into play in server resources and resource allocation. Are you using this server dedicated to only hosting one website or are you running other services on it as well?
Post up more info and I'm sure people will start chiming in with tips or ideas on how to make your site eat less resources.
Well, not an easy answer,
Well, not an easy answer, blucches. The site uses the only Drupal instance under its domain, but there are other Drupal installs (4.7 and 5.1) under other domains. All of them are sharing the httpd and the mysql. However, the server failures started just when we put this specific site into production.
Modules? Lots of them (kudos to the Drupal developer community):
addtofavorites
dbmaintenance
aggegator
comment
contact
forum
help
locale
menu
ping
profile
search
statistics
taxonomy
throttle
tracker
upload
image
image attach
image gallery
image assist
remember me
account reminder
adsense
akismet
authorship
forward
coolfilter
bbcode
google analytics
google sitemap
legal
login destination
login tobbogan
mailhandler
mailsave
meta tags
paging
path redirect
pathauto
printer fiendly
scheduler
service links
similar entries
simple news
site menu
switchtheme
taxonomy breadcrumb
taxonomy context
faxonomy force all
text link ads
trackback
userplus
adsense injector
feedburner
tagadelic
external links
As you can see, it's obviously difficult to go disabling them one by one and waiting until a server failure happens (I currently have around 6 failures a day, with all of them enabled). What I'm asking for is a way to monitor the server so we can know which module/feature caused the failure. What/where should we be looking for?
Thanks for any assistance
--
Albert