nginx error: (99: Cannot assign requested address)

问题:

I am running Ubuntu Hardy 8.04 and nginx 0.7.65, and when I try starting my nginx server:

$ sudo /etc/init.d/nginx start

I get the following error:

Starting nginx: [emerg]: bind() to IP failed (99: Cannot assign requested address)

where "IP" is a placeholder for my IP address. Does anybody know why that error might be happening? This is running on EC2.

My nginx.conf file looks like this:

user www-data www-data;
worker_processes  4;

events {
    worker_connections  1024;
}

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

    access_log /usr/local/nginx/logs/access.log;

    sendfile        on;
    tcp_nopush      on;
    tcp_nodelay         on;

    keepalive_timeout  3;

    gzip  on;
    gzip_comp_level 2;
    gzip_proxied any;
    gzip_types  text/plain text/css application/x-javascript text/xml application/xml
    application/xml+rss text/javascript;

    include /usr/local/nginx/sites-enabled/*;

}

and my /usr/local/nginx/sites-enabled/example.com looks like:

server {

        listen   IP:80;
        server_name  example.com;
        rewrite ^/(.*) https://example.com/$1 permanent;

       }

server {

        listen   IP:443 default ssl;

        ssl         on;
        ssl_certificate     /etc/ssl/certs/myssl.crt;
        ssl_certificate_key /etc/ssl/private/myssl.key;

        ssl_protocols       SSLv3 TLSv1;
        ssl_ciphers ALL:!ADH:RC4+RSA:+HIGH:+MEDIUM:-LOW:-SSLv2:-EXP;

        server_name example.com;

        access_log /home/example/example.com/log/access.log;
        error_log /home/example/example.com/log/error.log;

        }

回答1:

With Amazon EC2 and elastic IPs, the server doesn't actually know its IP as with most any other server.

So you need to tell your linux to allow processes to bind to the non-local address. Just add the following line into /etc/sysctl.conf file:

# allow processes to bind to the non-local address
# (necessary for apache/nginx in Amazon EC2)
net.ipv4.ip_nonlocal_bind = 1

and then reload your sysctl.conf by:

$ sysctl -p /etc/sysctl.conf

which will be fine on reboots.

回答2:

To avoid hard-coding the IP address in the config, do this:

listen *:80

回答3:

As kirpit mentioned above you'll want to allow linux processes to bind to a local IP address:

nano /etc/sysctl.conf

# allow processes to bind to the non-local address
net.ipv4.ip_nonlocal_bind = 1

sysctl -p /etc/sysctl.conf

Then you want to add the private ip address that is associated with your elastic ip and add that to your sites config:

nano /etc/nginx/sites-available/example.com

Reload nginx:

service nginx reload

All done!

回答4:

There might be remaining process/program that's using/listening at port 80.

You can check that using netstat -lp. Kill that process and start nginx.

回答5:

With Amazon EC2 and elastic IPs, the server doesn't actually know its IP as with most any other server. So in the apache virtual host files at least you put *:80 rather than your elastic ip :80

Then it works properly. So theoretically, doing *:80 for nginx should work the same but when you do you get [emerg]: bind() to 0.0.0.0:80 failed (98: Address already in use). Haven't found a solution yet. .

回答6:

For people who might be dealing with this in the future, I just looked up my private IP in the AWS instance and bound to that. I verified that nginx was able to listen publicly and perform my rewrite after that. I could not do *:PORT as I had an internal server I was proxying to.

回答7:

If you are using Network Manager, you have to wait to raise the network interface before starting the service:

systemctl enable NetworkManager-wait-online.service

回答8:

For Amazon EC2 and elastic IPs, sysctl.conf will not work as nginx still not listen on eth0.

So, you need to listen *;