Php-fpm is using all the cpu resources 90%


#1

Hi I have linode 2GB plan on Ubuntu 16.04 LTS - Nginx -Php5 and php-fpm is using all the cpu resources 90%.

in debian i hadn’t such a problem.

any ideas about how to solve?


#2

Have you tried to check your Google Analytics (traffic analysis tool) to see how much traffic you are receiving during this extreme cpu usage?

There could be several reasons for this. The key is to ask the questions what changed between now and then (when things were fine). Did you change your WP theme? Did you install a new plugin? Have you scanned your website? etc


#3

try a reboot to see if it’s a bug. it should not require all the cpu, even with a lot of users. EE is build for that


#4

sorry for late reply.

@Garikai_Dzoma hmm that could be hard. since there are approximately 10 websites on that. the things i changed was OS (debian to Ubuntu) moved to another VPS provider. And php cron jobs order. i am not sure if i made php cron jobs not to cross each other time lines.

i attached the graphs. i may miss some points but i didn’t see any relativity between cpu and network and disk operations. this is CPU Monthly graph.

this is network Monthly graph.

this is IPv6 Monthly graph.

And disk IO graphs.

@virtubox indeed. i did the reboot but the issue persists.

i was using debian OS before with the same sites and there weren’t such problems. moved the VPS to another provider and installed the sites on Ubuntu this time and the issues started to occur.


#5

Seems like a bug to me. Stick with Debian instead.


#6

I Wonder if there are any EE users with the same setup (linode vps + EE + Ubuntu) and experience the same situation though


#7

Hi, here digital ocean, ubuntu 16.04 from 4 days ago i started to get php-fpm kill my server, load avg goes from 1.5 to 101.5, and stop responding. Im test this happend when im clean w3tc clean caches.


#8

If your EE setup is hogging server resources either one or more of the following is happening:

  • your server is not powerful enough for your demand;
  • your server scripts are poorly written (what makes sense when you say the overload happens after cleaning the server cache), usual suspects are “stats” plugins, star ratings, and so on;
  • your blog is under some kind of attack (xmlrpc.php, brute force, or other).

You should take a look at your logs in order to find out what’s going on, what doesn’t void the need of deactivating all plugins (but Nginx Helper and cache) in order to see if the issue gets some relief or not.


#9

here is the latest update about this situation.

i migrated all the wordpress sites to debian config (The same ram, hdd, cpu.)

no resource problem since…


#10

I had a problem with digital ocean, but I discovered the problem. A plugin was raising high cpu peaks that went from 100%. After a few adjustments everything went back to normal.

https://content.screencast.com/users/FBSena/folders/Jing/media/3f88df7d-c27b-4916-8413-4f347ce8137c/2017-06-09_1320.png


#11

Would you mind telling the name of the plugin?


#12

Hello I had this problem with (All In One WP Security & Firewall).


#13

not enough ram, I’m change the provider and change php5.6 to php7 working stable!


#14