When nginx cannot connect to the your php fpm sock

Today was one of those days when I got an error on my play server out of nowhere. I hadn't changed much in the last few days, so I was a little confused about that. One of the websites was unable to work. I received a <a href="https://tools.ietf.org/html/rfc2616#section-10.5.3">502 Bad Gateway</a> error. The error log from Nginx showed the following: <code><code></code></code> <pre>2020/12/07 20:22:00 [crit] 1846#1846: 5 connect() to unix:/var/run/php/example.sock failed (2: No such file or directory) while connecting to upstream, client: 89..., server: example.com, request: "GET / HTTP/1.1", upstream: "fastcgi://unix:/var/run/php/website.sock:", host: "www.website.de" 2020/12/07 20:22:00 [crit] 1846#1846: 5 connect() to unix:/var/run/php/example.sock failed (2: No such file or directory) while connecting to upstream, client: 89..., server: example.com, request: "GET / HTTP/1.1", upstream: "fastcgi://unix:/var/run/php/example.sock:", host: "www.example.com" 2020/12/07 20:22:01 [crit] 1846#1846: 5 connect() to unix:/var/run/php/example.sock failed (2: No such file or directory) while connecting to upstream, client: 89..., server: example.com, request: "GET / HTTP/1.1", upstream: "fastcgi://unix:/var/run/php/example.sock:", host: "www.example.com" 2020/12/07 20:22:47 [crit] 1878#1878: 127 connect() to unix:/var/run/php/example.sock failed (2: No such file or directory) while connecting to upstream, client: 89..., server: example.com, request: "GET / HTTP/1.1", upstream: "fastcgi://unix:/var/run/php/example.sock:", host: "www.example.com" 2020/12/07 20:23:45 [crit] 1878#1878: 147 connect() to unix:/var/run/php/example.sock failed (2: No such file or directory) while connecting to upstream, client: 207..., server: example.com, request: "GET /2017-April-Marek-Konwa HTTP/1.1", upstream: "fastcgi://unix:/var/run/php/example.sock:", host: "www.example.com" 2020/12/07 20:24:24 [crit] 1878#1878: 288 connect() to unix:/var/run/php/example.sock failed (2: No such file or directory) while connecting to upstream, client: 36..., server: example.com, request: "GET /video-277.html?fbclid=IwAR2j2xofNyjzUh0hbnweMESDGhYIdC0qMn1IR5qDln8vktcedR-6LPv_k68 HTTP/1.1", upstream: "fastcgi://unix:/var/run/php/example.sock:", host: "www.example.com", referrer: "https://l.facebook.com/"</pre>

You like what you find here? A donation helps to keep this project running.

First, I checked the sockets and found ... nothing: <code><code></code></code> <pre>root@play:~# ls -lsa /var/run/php/ total 3 0 drwxr-xr-x 2 https://www.nginx.com/resources/wiki/start/topics/tutorials/debugging/#socket-leaks">https://www.nginx.com/resources/wiki/start/topics/tutorials/debugging/w-r--r-- 1 root root 4 Dec 7 20:27 php7.2-fpm.pid 0 srw-rw---- 1 https://www.nginx.com/resources/wiki/start/topics/tutorials/debugging/#socket-leaks">https://www.nginx.com/resources/wiki/start/topics/tutorials/debugging/t;https://www.nginx.com/resources/wiki/start/topics/tutorials/debugging/#socket-leaks">https://www.nginx.com/resources/wiki/start/topics/tutorials/debugging/#socket-leaks</a>), I rethought my changes in the last few days. I had just added a subdomain to example.com (subdomain dev.example.com). I isolated the dev with their own PHP pool. The directions and permissions were correct. While I reviewed the pool configs, I realized that I had named the pools the same. dev.example.com AND example.com were configured under [example]. And that was the problem. PHP was still able to start, but instead of throwing an error, the two configs (with different sockets) were running under the newer sock (dev.example.sock). The fix: just change the name of the pool (now I had two pools [dev.example] and [example]) and restart the PHP service. That's it.