-1

It looks like an application specific issue and my server(Linode) is using an excessive amount of I/O.

(config: Ubuntu 12.0.4+Nginx+MariaDB+PHP-FPM+APC+iONCube+igBinary)

I'm trying to figure out where is right direction but it can be a very complex issue to debug. However my I/O will certainly cause a problem for your system.

Total DISK READ: 0.00 B/s | Total DISK WRITE: 132.51 M/s 
TID PRIO USER DISK READ DISK WRITE SWAPIN IO> COMMAND 
26474 be/4 www-data 0.00 B/s 91.29 M/s 0.00 % 0.31 % php-fpm: pool www 
26255 be/4 www-data 0.00 B/s 72.49 M/s 0.00 % 0.82 % php-fpm: pool www 

The problem seems to be specifically with PHP based on your error logs:

: 66.249.76.233, server: myprestashop.com, request: "POST /module/mailalerts/actions?process=check HTTP/1.1", upstream: "fastcgi://unix:/tmp/php5-fpm.sock:", host: "myprestashop.com", referrer: "https://myprestashop.com/farmacia-y-medicinas/15038-garamicina-frasco-ampula-5-80mg-antibiotico-7501050610537.html"
2014/05/03 04:27:32 [crit] 13627#0: *25936 connect() to unix:/tmp/php5-fpm.sock failed (2: No such file or directory) while connecting to upstream, client: 66.249.76.233, server: myprestashop.com, request: "GET /sistema-nervioso/11634-rapix-capsulas-10-10mg-7501314703821.html HTTP/1.1", upstream: "fastcgi://unix:/tmp/php5-fpm.sock:", host: "myprestashop.com"

I restarted PHP:

root@luna:/etc/nginx/sites-enabled# killall php5-fpm root@luna:/etc/nginx/sites-enabled# /etc/init.d/php5-fpm start

However after that I get new messages in the error log:

2014/05/03 04:31:43 [error] 13627#0: *26155 FastCGI sent in stderr: "PHP message: PHP Warning: array_shift() expects parameter 1 to be array, boolean given in /sites/myprestashop.com/public/modules/productcomments/ProductComment.php on line 140 
PHP message: PHP Warning: array_shift() expects parameter 1 to be array, boolean given in /sites/myprestashop.com/public/modules/productcomments/ProductComment.php on line 140" while reading response header from upstream, client: 65.55.55.231, server: myprestashop.com, request: "GET /higiene-personal/4519-koleston-cast-osc-viol-int-366-7501007455174.html HTTP/1.1", upstream: "fastcgi://unix:/tmp/php5-fpm.sock:", host: "myprestashop.com" 
2014/05/03 04:32:06 [error] 13626#0: *26195 FastCGI sent in stderr: "PHP message: PHP Notice: Undefined index: product_mode in /sites/myprestashop.com/public/cache/smarty/compile/b0/9a/32/b09a3295da69c6626cf2789ed62cd44f2fbed7f6.file.product-list.tpl.php on line 36 
PHP message: PHP Notice: Undefined index: product_mode in /sites/myprestashop.com/public/cache/smarty/compile/b0/9a/32/b09a3295da69c6626cf2789ed62cd44f2fbed7f6.file.product-list.tpl.php on line 146 
PHP message: PHP Notice: Undefined index: product_mode in /sites/myprestashop.com/public/cache/smarty/compile/b0/9a/32/b09a3295da69c6626cf2789ed62cd44f2fbed7f6.file.product-list.tpl.php on line 146" while reading response header from upstream, client: 66.249.76.233, server: myprestashop.com, request: "GET /26-cuidado-del-cabello?amp%252525252525253Bselected_filters=%3Fcontroller%3D404&p=14 HTTP/1.1", upstream: "fastcgi://unix:/tmp/php5-fpm.sock:", host: "myprestashop.com"

It looks like the code itself has issues that will need to be resolved.

EDIT

there are 2 APC instance, one in modes-available and the other one in conf.d

also php-fpm log

[03-May-2014 11:38:11] ERROR: An another FPM instance seems to already listen on /tmp/php5-fpm.sock
[03-May-2014 11:38:11] ERROR: FPM initialization failed

may be the solution here is to remove php+igbinary+ioncube

1
  • 2
    1) Who are you talking to ("problem for your system", "based on your error logs")? 2) Is there a question here?
    – MadHatter
    May 3, 2014 at 10:47

1 Answer 1

0

Every time I've seen something like this, it has been because of

  • Apache configuration / .htaccess has defined something that leads to infinite loop -> AIEEEE
  • Someone created a PHP or whatever script which had some kind of for loop doing something nasty or otherwise was a source of problems
  • Some script tried to call some external resource but could not do that, and somehow decided to retry too fast -> AIEE
  • Your PHP script has some logging enabled and under X conditions it starts to flood the logs -> AIEEEE

Any of these possible?

1
  • I'm noticing that APC is increasing the the CPU levels to 500%!!! I will update my original post with APC data an php log
    – Locke
    May 3, 2014 at 16:36

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .