I’m facing the following problem on my nginx/php-fpm server : I get random 502 (bad gateway) errors on AJAX calls (autocompletion actually). Everything else is working fine, except autocompletion.
On my php-fpm.log, I got a huge amount of process dying (signal 11) after like 18 hours or so… This doesn’t look correct right?
What can make this happen? I already tweaked pm.max_request, max execution time… And I’m having a dynamic php-fpm pool.
A good starting point is to edit php.ini and add/edit this line from:
; error_log = filename
error_log = /var/log/php.errors
be sure the log file is owned by the nginx user.