Nearly two weeks I'm stuck with this LEMP problem in which I get only one error anywhere and I don't feel this error helps me to solve my problem:
*1 connect() failed (111: Connection refused) while connecting to upstream,
client: MY_IP_ADDRESS, server: example.com, request: "GET /
HTTP/2.0", upstream: "fastcgi://127.0.0.1:9000".
nginx -t
, access.log
, and /etc/init.d/php*-fpm status
returns no additional errors.
It's not the first time I get a "vague" connection refused
error from Nginx while no other error log indicates what might be the source of the problem:
How to further debug Nginx in such circumstances?
BTW, I tried to contact the Nginx developers on this but they have shutted down any reasonable option to contact them (for example, they don't get emails from Gmail senders or GitHub issues).
Best Answer
Your
nginx
config specifies that it is supposed to connect toPHP-FPM
using TCP to127.0.0.1:9000
but yourPHP-FPM
config specifies that it doesn't listen on TCP port 9000 but instead uses unix socket/run/php/php7.0-fpm.sock
so you getconnection refused
.