在 lxc 容器内由 PM2 管理的 NodeJS 服务器的 502 错误网关

Question

I have a digital ocean droplet running Ubuntu 18.04 and inside is is an lxc container. I have two applications in that container.

The first application (a client) lives at /var/www/html and the second one is the NodeJS application that lives at /var/www/my-site/ . The Node application inside the container is managed by pm2 and everything seems to be working fine thus far because when I type in curl http://localhost:3000 at the container terminal, I get back the desired output.

Inside the main droplet (not the container) under /etc/nginx/sites-available , I have the following two server blocks - default and my-site .

The first app works fine when I try to access it through the browser via my domain but the NodeJS application returns a 502 Bad Gateway when I try to access it through sub.mydomain.com . pm2 start inside the container tells me that the node application status is online .

Here is my default server block file. This works. When I visit mydomain.com, my site shows up fine.

# HTTP — redirect all traffic to HTTPS
server {
    listen 80;
    listen [::]:80 default_server ipv6only=on;
    return 301 https://$host$request_uri;
}

server {
    # Enable HTTP/2
    listen 443 ssl http2;
    listen [::]:443 ssl http2;
    server_name mydomain.com;
    # Use the Let’s Encrypt certificates
    ssl_certificate /etc/letsencrypt/live/mydomain.com/fullchain.pem;
    ssl_certificate_key /etc/letsencrypt/live/mydomain.com/privkey.pem;
    # Include the SSL configuration from cipherli.st
    include snippets/ssl-params.conf;

    root /var/www/html;

    index index.html index.htm index.nginx-debian.html;

    location / {
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header  X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_pass http://container_ip_address /;
    }
}

Now here is the other server block - my-site .

# Upstream config
upstream site_upstream {
    server 127.0.0.1:3000;
    keepalive 64;
}

server {
    # Enable HTTP/2
    listen 443 ssl http2;
    listen [::]:443 ssl http2;
    server_name sub.mydomain.com www.sub.mydomain.com;
    root /var/www/my-site;

    # Use the Let’s Encrypt certificates
    ssl_certificate /etc/letsencrypt/live/mydomain.com/fullchain.pem;
    ssl_certificate_key /etc/letsencrypt/live/mydomain.com/privkey.pem;
    # Include the SSL configuration from cipherli.st
    include snippets/ssl-params.conf;

    location / {
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-NginX-Proxy true;
        proxy_pass http://site_upstream;
        proxy_ssl_session_reuse off;
        proxy_set_header Host $http_host;
        proxy_cache_bypass $http_upgrade;
        proxy_redirect off;
    }
}

I have set the A Record for my subdomain on my domain's DNS settings, to my droplet's IP address and I have also created a symbolic link to /etc/nginx/sites-enabled for the my-site server block.

I have scoured the internet for a solution to this problem but nothing seems to be working. What am I missing?

Your help would be greatly appreciated. Thanks.

Answer 1

The problem here was that requests to the sub domain were not being directed to the lxc container. I solved this by adding the following inside the my-site server block.

location / {
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header  X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_pass http://container_ip/;
    }

After that I added an asterisk to the next location block.

location /* {
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-NginX-Proxy true;
        proxy_pass http://site_upstream;
        proxy_ssl_session_reuse off;
        proxy_set_header Host $http_host;
        proxy_cache_bypass $http_upgrade;
        proxy_redirect off;
    }

Another way of getting around this issue was by including the sub-domain in the server_name directive for the default server block. This worked but the only problem was that nginx would complain that it had to ignore the server I had set up in the my-site server block when you ran nginx -t , otherwise, it worked just fine.