Docker Swarm - PHP-FPM not working with Nginx running on manager - php

Question

I've build a cluster (1 manager, 4 workers). Only the manager has a public IP address, workers and the manager are on a private network.
I tried to build a webserver (Nginx + PHP-FPM) with Docker Swarm. So I set the Nginx container on the manager, so I can request it from outside the private network. If I do that, the container get a upstream timed out (110: Connection timed out) while connecting to upstream error while requesting a PHP file. If I run it on a worker node, everything works fine, but the Nginx isn't accessible from outside the private network (I can only request with curl on the manager or every worker).
Do you guys have any idea ? I really don't understand why running Nginx on the manager makes PHP-FPM timed out. Thanks !
Here the docker-compose.yml file:
version: '3.8'
services:
nginx:
image: arm32v5/nginx:latest
container_name: webserver_nginx
ports:
- 80:80
volumes:
- /media/storage/webserver/nginx/nginx.conf:/etc/nginx/nginx.conf
- /media/storage/webserver/nginx/log:/var/log/nginx
- /media/storage/www:/var/www
links:
- php
networks:
- webserver
deploy:
placement:
constraints:
- "node.role==manager"
php:
image: arm32v5/php:7.4-fpm
container_name: webserver_php
volumes:
- /media/storage/www:/var/www
- /media/storage/webserver/nginx/www.conf:/usr/local/etc/php-fpm.d/www.conf
networks:
- webserver
links:
- nginx
ports:
- 9000:9000
networks:
webserver:
driver: overlay
attachable: true
Nginx configuration:
user nginx;
worker_processes 1;
error_log /var/log/nginx/error.log warn;
pid /var/run/nginx.pid;
events {
worker_connections 1024;
}
http {
include /etc/nginx/mime.types;
default_type application/octet-stream;
log_format main '$remote_addr - $remote_user [$time_local] "$request" $status $body_bytes_sent "$http_referer" "$http_user_agent" "$http_x_forwarded_for"';
access_log /var/log/nginx/access.log main;
sendfile on;
keepalive_timeout 65;
server_tokens off;
server {
listen 80;
error_page 500 502 503 504 /50x.html;
location / {
index index.php index.html index.htm;
root /var/www/;
}
location ~ \.php$ {
root /var/www/;
fastcgi_split_path_info ^(.+\.php)(/.+)$;
fastcgi_pass php:9000;
fastcgi_index index.php;
include fastcgi_params;
fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
fastcgi_param PATH_INFO $fastcgi_path_info;
}
}
}

Answer 1

I had the similar kind of issue and the solution was to use the docker service name in upstream.
Here is my setup
3 EC2 with Docker Swarm
1 Swarm Manager ( Nginx proxy deployed )
2 Swarm Worker (Frontend + backend deployed)
Flow
Browser ->> Nginx-proxy ->> React-frontend (Nginx) ->> Go-backend
nginx.conf
http {
upstream allbackend {
server exam105_frontend:8080; #FrontEnd Service Name
}
server {
listen 80;
location / {
proxy_pass http://allbackend;
}
}
}
Docker Swarm Services
AWS Note:
Remember to open port 80 publicly and 8080 (or whatever port number you are using) for internal communication in Security Group of your AWS setup, otherwise you won't be able to access the service.