200 connect() to unix:/tmp/uwsgi.sock failed

Question

I am configuring nginx with uwsgi on EC2, I have check logs in file /var/log/nginx/error.log.

I am getting this error:

200 connect() to unix:/tmp/uwsgi.sock failed (2: No such file or directory) while connecting to upstream

My uwsgi.sock location is /var/www/myapp/

How can I change the file location from uwsgi://unix:/tmp/uwsgi.sock to uwsgi://unix:/var/www/myapp/ in configuartion?

You should specify the filename where the .sock file resides. If your socket file is in the directory `/var/www/myapp/`, your config should be `/var/www/myapp/uwsgi.sock` — Freek Wiekmeijer, Jun 15 '15 at 08:55
Thanks for your reply but in nginx.conf file I am providing the path /var/www/myapp/uwsgi.sock. Still getting the error. — Abhishek, Jun 15 '15 at 12:27
Under which user does uWSGI run and does that user have write access to /var/www/myapp ? — Freek Wiekmeijer, Jun 15 '15 at 12:51
user is ubuntu and /var/www/myapp has the write permisssion. I have set the permission 666 for this. — Abhishek, Jun 15 '15 at 16:39
Does it create the .sock file if you simply type uwsgi -s /var/www/myapp/uwsgi.sock on the terminal (as user ubuntu)? — Freek Wiekmeijer, Jun 15 '15 at 16:40
and when you run it manually with all the config options on the command line (or with a reference to uwsgi.conf)? — Freek Wiekmeijer, Jun 15 '15 at 17:13
What is the full error_log from nginx? Does it mention any sort of duplicates of anything? ;) — cnst, Jun 28 '16 at 18:54

score 2 · Accepted Answer · edited May 23 '17 at 11:51

This might be a problem with classical namespaced Temp, please see this: http://fedoraproject.org/wiki/Features/ServicesPrivateTmp

but you mentioned that you have set your sock location to your app directory, have you done it in your nginx configuration for that virtual host? you will definitely need to restart your web server for that changes to take affect.

Please refer this question and its answer, which might be useful in this case. Got 'No such file or directory' error while configuring nginx and uwsgi

Please refer to the section where in nginx.conf file, uwsgi_pass is mentioned. if you have not made changes like that, please do so and restart the webserver. I think that must solve the problem.

score -2 · Answer 2 · answered Jul 02 '16 at 00:00

You have not responded to my comment yet:

What is the full error_log from nginx? Does it mention any sort of duplicates of anything? ;) – cnst Jun 28 at 18:54

The reason I'm asking is that the default configuration on Debian, with the sites-available/ and sites-enabled/ directories, has an include that does not ignore strategically-placed backup files (like those ending in ~).

I once wasted a good amount of time trying to "quickly" do some PoC, only being puzzled why things didn't work which I know 100% should totally work, and why does my error_log report some mysterious bind(2) issues:

23:46:38 [emerg] 25509#0: bind() to [::]:80 failed (98: Address already in use)

OP has already mentioned what his log file says. had it been the case of already used address, the upstream connection error would not have shown in the first place. — openrijal, Jul 04 '16 at 09:55

200 connect() to unix:/tmp/uwsgi.sock failed

2 Answers2