Hi everyone...

I noticed after one of the recent upgrades to the 2.2.x "RELEASE" series
everything works perfectly fine for a while but then, I get "502 Bad Gateway"
message when I attempt to access the web GUI.

Up until recently, I had been "fixing" this by physically powering off the
firewall by pushing the power button which causes a clean shutdown. Then I
would power it back up and it works fine for a random amount of time - usually
days or weeks, sometimes even months - but this last time it only lasted about
23 hours.

I am currently running the 2.3.1-RELEASE-p5 (amd64) nanobsd (4g) distribution.

Today, I have enabled the ssh service and checked the nginx*.log files and I 
see:

[2.3.1-RELEASE][[email protected]]: clog -f /var/log/nginx-error.log

----8<----
2016/07/05 12:40:01 [error] 48883#0: *257237 upstream timed out (60: Operation
timed out) while reading response header from upstream, client: 192.168.254.4,
server: , request: "GET /getstats.php HTTP/1.1", upstream:
"fastcgi://unix:/var/run/php-fpm.socket", host: "vai.revpol.com:4443",
referrer: "https://vai.revpol.com:4443/";
clog: ERROR: could not write output (Bad address)
----8<----


At that same time, system.log shows that the php-fpm.socket socket does not 
exist:
----8<----
Jul 5 12:47:00 vai vai.revpol.com nginx: 2016/07/05 12:47:00 [crit] 48883#0:
*257442 connect() to unix:/var/run/php-fpm.socket failed (2: No such file or
directory) while connecting to upstream, client: 192.168.254.4, server: ,
request: "GET /getstats.php HTTP/1.1", upstream:
"fastcgi://unix:/var/run/php-fpm.socket:", host: "vai.revpol.com:4443",
referrer: "https://vai.revpol.com:4443/";
----8<----


>From the console menu, I can choose option 16 (Restart PHP-FPM), and then the
web gui is accessible again.

Enter an option: 16
----8<----
>>> Killing php-fpm
>>> Starting php-fpm
*** Welcome to pfSense 2.3.1-RELEASE-p5 (amd64 nanobsd) on vai ***
----8<----

So, I am suspecting that the php-fpm process is dying (forgot to run a ps
command before restarting it).

Right now, /tmp/php_errors.txt is a zero byte file but I suspect that may be
due to the restart of php-fpm due to its timestamp. I will take a look at this
file the next time the gui dies.


Is there anything I can do to increase debugging to help identify why this
process is dying?

Additional info:  Typically I have a Firefox tab "idling" on the dashboard
page which includes the "Traffic Graphs" widget with 4 graphs,
Autoscale=Follow and 1 second updates.


Thanks!

Bill

-- 
Bill Arlofski
Reverse Polarity, LLC
http://www.revpol.com/blogs/waa
-- Not responsible for anything below this line --
_______________________________________________
pfSense mailing list
https://lists.pfsense.org/mailman/listinfo/list
Support the project with Gold! https://pfsense.org/gold

Reply via email to