Server keeps going down with 502 Bad Gateway error

Events happening in the community are now at Drupal community events on www.drupal.org.
r2b2's picture

Hi,

We recently moved to DigitalOcean and used nginx. My colleague who set it up left the company just right after we launch the site and so we are at a lost on how to solve this problem.

Here are the server logs, if it may help. There are also several config files, which one do I need to display here?

I hope you can help us,thank you in advance.

tail -f /var/log/php5-fpm.log

[22-Apr-2013 08:00:33] NOTICE: [pool www] child 4923 started
[22-Apr-2013 08:23:25] NOTICE: [pool www] child 1569 exited with code 0 after 10287.032182 seconds from start
[22-Apr-2013 08:23:25] NOTICE: [pool www] child 5067 started
[22-Apr-2013 08:45:23] NOTICE: [pool www] child 1503 exited with code 0 after 14606.752794 seconds from start
[22-Apr-2013 08:45:23] NOTICE: [pool www] child 5212 started
[22-Apr-2013 08:56:12] NOTICE: Finishing ...
[22-Apr-2013 08:56:12] NOTICE: exiting, bye-bye!
[22-Apr-2013 08:56:13] NOTICE: fpm is running, pid 5441
[22-Apr-2013 08:56:13] NOTICE: ready to handle connections
[22-Apr-2013 08:58:11] WARNING: [pool www] server reached pm.max_children setting (10), consider raising it

Comments

This was result #2 from some

bhosmer's picture

This was result #2 from some googling of your pm.max_children error:

http://serverfault.com/questions/295766/what-is-wrong-in-my-php-fpm-conf...

The post seems to suggest splitting your setting in half.

more details

fully_stoned's picture

This is not related with Nginx, still if you give more information about your Ram, CPU cores/threads, your php-fpm settings then it'd be easier to suggest what you have been missing.

The answer is there

perusio's picture

you need to increase the maximum number of children and/or the maximum number of requests each worker accepts. Paste your php-fpm pool config here.

Ignore such messages

trimboot's picture

Hello.
You can ignore such messages
...
[22-Apr-2013 08:23:25] NOTICE: [pool www] child 1569 exited with code 0 after 10287.032182 seconds from start
[22-Apr-2013 08:23:25] NOTICE: [pool www] child 5067 started
...
because it's a NOTICEs anyway.
They are appear because your colleague (who left the company) have set the pm.max_request in pool configuration file
/etc/php5/fpm/pool.d/www.conf
It's the number of requests each child process should execute before respawning. Default value is "0".

Nginx

Group organizers

Group notifications

This group offers an RSS feed. Or subscribe to these personalized, sitewide feeds: