0

I have been having ongoing 504 errors (503s). It has gotten to the point where it is so bad everytime someone trys an update in wordpress they get a 504.

When troubleshooting I tried to adjust a timeout in php-fpm and the system crashes. I restored the file back to what it was and now I cannot get php5-fpm to load. Everytime I try a restart, reload, stop, start the system freezes and I have to force quit the VM in virt manager. I have been running in circles unable to find anything close to a solution.

These are the error logs I get:

tail -f /usr/local/var/log/nginx/error.log

Specifically:

2015/10/11 19:46:53 [error] 2429#0: *1779 upstream timed out (110: Connection timed out) while reading response header from upstream, client: 10.30.1.119, server: [redacted], request: "GET /robots.txt HTTP/1.1", upstream: "fastcgi://unix:/var/run/php5-fpm-www-1.sock:", host: "10.30.1.110"

My php-fpm.conf file is configured as such:

pid = /var/run/php5-fpm.pid
error_log = /var/log/php5-fpm.log
include=/etc/php5/fpm/pool.d/*.conf

The rest is commented out.

I do not even see a log file when I follow the path.

I am not sure if any of this is related to the 504 error. We have in the past experienced a complete freeze CPU spike which required the system to be force restarted. I am not sure how we solved it before. Eventually something worked.

Is there something I should be checking that I am not?

Carl Bennett
  • 783
  • 3
  • 7
  • 14
renon
  • 1
  • 2
  • Have you tried restarting nginx? `sudo nginx -s reload` - I think nginx might have cached the upstream server's hostname and the IP may have changed. – Carl Bennett Oct 15 '15 at 00:08
  • Hello Carl, yes I am able to restart (reload) nginx without issue. Its whenever I try to do anything with frpm-php that I have to power down the machine. I did read that there is a known issue with this. Things were working fine until I tried to solve the issue. All I did though is create an override script, which worked. I was then able to restart fpm-php. however, it froze again, I then powered it down and the issue came back. So I deleted my script. Are you aware of any logs I should be checking to help with this. Other than what I listed above. I am new to nginx. Thanks! – renon Oct 15 '15 at 04:40
  • I am still getting this issue. If I relaod the cPU starts going insane and spiking on intervals. I was able to log in to the cluster machine and do a ps aux. A process called kipmi0 is taking up 40% of the cpu. – renon Oct 18 '15 at 01:58
  • checking the syslog says apparmor DENIED. run-shm-lyyng-ust-wait-5 – renon Oct 18 '15 at 02:04
  • I enable php-fpm debugging and I get this error message: Error pid 2920 An another FPM instance seems to already listen on /var/run/php5-fpm-www-1.sock – renon Oct 19 '15 at 04:00
  • I get the above error when running: # php5-fpm -y /etc/php5/fpm/php-fpm.conf – renon Oct 19 '15 at 04:06

0 Answers0