ipv4 & ipv6 Virtual Hosting - address in use

[Maxim Dounin]

Hello!

On Mon, Dec 27, 2010 at 10:55:35AM -0500, petteyg359 wrote:

> Linux 2.6.34-gentoo-r12 x86_64 Intel(R) Core(TM) i7 CPU 930 @ 2.80GHz
> GenuineIntel GNU/Linux
>  # nginx -V
> nginx version: nginx/0.8.53
> TLS SNI support enabled
> configure arguments: --prefix=/usr --sbin-path=/usr/sbin/nginx
> --conf-path=/etc/nginx/nginx.conf
> --error-log-path=/var/log/nginx/error_log --pid-path=/var/run/nginx.pid
> --lock-path=/var/lock/nginx.lock --user=nginx --group=nginx
> --with-cc-opt=-I/usr/include --with-ld-opt=-L/usr/lib
> --http-log-path=/var/log/nginx/access_log
> --http-client-body-temp-path=/var/tmp/nginx/client
> --http-proxy-temp-path=/var/tmp/nginx/proxy
> --http-fastcgi-temp-path=/var/tmp/nginx/fastcgi
> --http-scgi-temp-path=/var/tmp/nginx/scgi
> --http-uwsgi-temp-path=/var/tmp/nginx/uwsgi --with-ipv6 --with-pcre
> --without-http_autoindex_module --without-http_browser_module
> --without-http_charset_module --without-http_geo_module
> --without-http_map_module --without-http_memcached_module
> --without-http_referer_module --without-http_ssi_module
> --without-http_split_clients_module --without-http_userid_module
> --with-http_realip_module --with-http_ssl_module
> --without-mail_imap_module --without-mail_pop3_module
> --without-mail_smtp_module
> 
> server {
> listen [::]:80 default_server;
> listen 80 default_server;
> return 444;
> }
> #According to docs, this config should fail, because without
> ipv6only=on,
> #this line should automatically listen on ipv4 interfaces also.

This shouldn't fail as long as you have ipv6-only bind as
default in your system.  It won't even fail on OSes where binds on 
ipv6 with ipv4 doesn't conflict with ipv4 explicit binds.  This is 
unrelated though.

>  # netstat -lp | grep nginx
> tcp        0      0 *:http                  *:*                    
> LISTEN      31704/nginx.conf
> tcp6       0      0 [::]:http               [::]:*                 
> LISTEN      31704/nginx.conf
>
> server {
> listen [::]:80 default_server ipv6only=on;
> return 444;
> }
> #I remove the listen 80 line, and add ipv6only=on
>  # /etc/init.d/nginx start
>  * Checking nginx' configuration ...
> the configuration file /etc/nginx/nginx.conf syntax is ok
> configuration file /etc/nginx/nginx.conf test is successful             
>                                                                         
>                                                        [ ok ]
>  * Starting nginx ...
> [emerg]: bind() to [2a01:4f8:130:9101::3]:80 failed (98: Address already
> in use)
> [emerg]: bind() to [2a01:4f8:130:9101::3]:80 failed (98: Address already
> in use)
> [emerg]: bind() to [2a01:4f8:130:9101::3]:80 failed (98: Address already
> in use)
> [emerg]: bind() to [2a01:4f8:130:9101::3]:80 failed (98: Address already
> in use)
> [emerg]: bind() to [2a01:4f8:130:9101::3]:80 failed (98: Address already
> in use)
> [emerg]: still could not bind()
>  * start-stop-daemon: failed to start `/usr/sbin/nginx'
>  * Failed to start nginx                                                
>                                                                         
>                                                        [ !! ]
>  * ERROR: nginx failed to start
> 
> Not only does it not work as described, it seems to be trying to bind a
> specific address multiple times, and maybe succeeding the first time,
> because there's nothing else running on any interface port 80, so it
> fails since it can't successfully bind the same address a second time.

Could you please provide *full config* which triggers the problem?  
In particular, full list of server{} blocks with their "listen" 
directives.

Maxim Dounin