Forums

Error code: 502-backend

Is there any reason I got kicked off the system as user and developer for over 10 minutes? It seems there are many reasons why 'Error code: 502-backend' can happen but how do I identify the specific one in this case other than...

"wsgi.py has been touched"

2016-09-09 20:05:43 Fri Sep  9 20:05:43 2016 - *** /var/www/phadmin_pythonanywhere_com_wsgi.py has been touched... grace them all !!! ***
2016-09-09 20:05:43 ...gracefully killing workers...
2016-09-09 20:05:43 Gracefully killing worker 2 (pid: 3843)...
2016-09-09 20:05:56 Gracefully killing worker 1 (pid: 3839)...
2016-09-09 20:05:56 Gracefully killing worker 3 (pid: 3847)...
2016-09-09 20:07:46 worker 1 buried after 2 seconds
2016-09-09 20:07:46 worker 2 buried after 2 seconds
2016-09-09 20:07:46 worker 3 buried after 2 seconds
2016-09-09 20:07:46 uWSGI: GAME OVER (insert coin)
2016-09-09 20:07:46 chdir(): No such file or directory [core/uwsgi.c line 1541]
2016-09-09 20:07:46 chdir(): Stale file handle [core/uwsgi.c line 1561]
2016-09-09 20:07:46 VACUUM: unix socket /var/sockets/phadmin.pythonanywhere.com/socket removed.
2016-09-09 20:16:52 *** Starting uWSGI 2.0.5.1 (64bit) on [Fri Sep  9 20:14:59 2016] ***
phadmin | 40 posts | Sept. 10, 2016, 3:52 a.m. | permalink

Hi phadmin,

"wsgi.py has been touch" means that you edited/saved /var/www/phadmin_pythonanywhere_com_wsgi.py

When we detect that, we automatically restart your webapp for you.

Do you mean that your webapp took 10 minutes to restart? Or do you meant that you were unable to access pythonanywhere for 10 minutes?

Staff conrad | 4232 posts | PythonAnywhere staff | Sept. 10, 2016, 4:32 p.m. | permalink

This time I took notes so re what was happening
so with regard to pythonanywhere.com access 

> TAB                  RESULT  
> Consoles             Was Available    
> Files                Crashed   
> Web                  Crashed  
> Schedule             Crashed   
> Databases            Crashed  
> 
> Send feedback     Opened text input screen

At which point it seemed access was being restored so stopped checking. I am pretty sure the forums link crashed as well but alas I didn't write it down so not 100% sure now.

Will have to look at the logs to see what I was doing if anything that might have caused crash.

phadmin | 40 posts | Sept. 10, 2016, 6:28 p.m. | permalink

Hi phadmin,

I wasn't sure about the timing before, but from your description, I believe that may have corresponded with one of our server outages for ~5-10min then. Don't think this would have been something that you did. Sorry about that!

Staff conrad | 4232 posts | PythonAnywhere staff | Sept. 11, 2016, 10:20 a.m. | permalink