Load balancing options

I'm personally not familiar with haproxy, but I know it's dead-simple to do load balancing with Nginx.

First create a 'custom app listening on port' application to reserve a port for your Nginx load balancer. For the rest of this example, I'll assume that this app is named "load_balancer" and its been assigned port 54321. Go ahead and assign that app to a website in the panel as well.

After you've created the app in the control panel, SSH into your server and run the following commands to build Nginx:

mkdir ~/src ~/tmp
export TMPDIR=~/tmp
cd ~/src
wget http://nginx.org/download/nginx-1.8.0.tar.gz
tar zxf nginx-1.8.0.tar.gz 
cd nginx-1.8.0
./configure --prefix=$HOME/webapps/load_balancer

Next, edit your `~/webapps/load_balancer/conf/nginx.conf to set the port and configure the backend servers that it's going to balance:

worker_processes  1;

events {
    worker_connections  1024;
}


http {
    include       mime.types;
    default_type  application/octet-stream;

    access_log  /home/testweb477/logs/user/load_balancer_access.log;
    error_log  /home/testweb477/logs/user/load_balancer_error.log  warn;

    sendfile        on;
    keepalive_timeout  65;

    upstream backend {
        server server1.domain.com:43210;
        server server2.domain.com:32145;
    }

    server {
        listen       54321;
        server_name  localhost;

        location / {
            proxy_pass http://backend;
        }

        error_page   500 502 503 504  /50x.html;
        location = /50x.html {
            root   html;
        }
    }
}

Finally, start your Nginx server:

~/webapps/load_balancer/sbin/nginx

At that point, the site to which you assigned the load_balancer application will be proxying requests to server1.domain.com and server2.domain.com, round-robin style. If you want it to handle traffic more-intelligently, you can use the any of the various load-balancing methods described in the Nginx load-balancing guide. For example, if you want to use the least_conn method (so requests fo to the server with the least number of active connections), you would modify the upstream block in the config so that it looks like this:

    upstream backend {
        least_conn;
        server server1.domain.com:43210;
        server server2.domain.com:32145;
    }

Also, note that the backend server configuration above includes the port numbers (random in this example) for the other servers. Be sure to use the "open port" option for those apps on the other servers.

Just remember to restart or reload your Nginx whenever you modify the config.

I'm sure you could do something similar with haproxy, if that's your tool-of-choice. I don't know the exact procedure, butt the core concepts should be the same:

• Create a 'custom app listening on port' application to reserve a port for haproxy
• Assign that app to a site
• Build haproxy from source and configure it listen on the assigned port and proxy traffic to your backend servers.

Hope that helps!