Disable IPv6 in nginx proxy_pass

Question

My server doesn't have IPv6 adresses.

However, when I use Nginx proxy_pass to upstream with IPv4 and IPv6, sometimes it tries to send outgoing requests using IPv6:

2013/07/30 00:25:06 [error] 1930#0: *1482670 connect() to [AAAA:BBBB:C:DDD:E:F:GGG:HHH]:443 failed (101: Network is unreachable) while connecting to upstream, client: AA.BB.CC.DD, server: example.com, request: "GET /download/file HTTP/1.0", upstream: "https://[AAAA:BBBB:C:DDD:E:F:GGG:HHH]:443/download/file", host: "example.com"

How can I disable IPv6 for outgoing requests in proxy_pass?

nginx.conf:

upstream download {
  server download.example.com:443;
  keepalive 8;
}
location /download {
  proxy_set_header      X-Forwarded-For  $proxy_add_x_forwarded_for;
  proxy_set_header      Connection "";
  proxy_ignore_headers  X-Accel-Redirect;
  proxy_http_version    1.1;
  resolver              8.8.8.8;
  resolver_timeout      5s;
  proxy_pass            https://download;
}

nginx -V:

nginx version: nginx/1.4.2
built by gcc 4.7.2 (Debian 4.7.2-5)
TLS SNI support enabled
configure arguments: --prefix=/etc/nginx --sbin-path=/usr/sbin/nginx --conf-path=/etc/nginx/nginx.conf --error-log-path=/var/log/nginx/error.log --http-log-path=/var/log/nginx/access.log --pid-path=/var/run/nginx.pid --lock-path=/var/run/nginx.lock --http-client-body-temp-path=/var/cache/nginx/client_temp --http-proxy-temp-path=/var/cache/nginx/proxy_temp --http-fastcgi-temp-path=/var/cache/nginx/fastcgi_temp --http-uwsgi-temp-path=/var/cache/nginx/uwsgi_temp --http-scgi-temp-path=/var/cache/nginx/scgi_temp --user=nginx --group=nginx --with-http_ssl_module --with-http_realip_module --with-http_addition_module --with-http_sub_module --with-http_dav_module --with-http_flv_module --with-http_mp4_module --with-http_gunzip_module --with-http_gzip_static_module --with-http_random_index_module --with-http_secure_link_module --with-http_stub_status_module --with-mail --with-mail_ssl_module --with-file-aio --with-http_spdy_module --with-cc-opt='-g -O2 -fstack-protector --param=ssp-buffer-size=4 -Wformat -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2' --with-ld-opt=-Wl,-z,relro --with-ipv6

OS: Debian Wheezy

Linux 3.2.0-4-amd64 #1 SMP Debian 3.2.46-1 x86_64 GNU/Linux

ip a

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 6c:62:6d:7a:ea:af brd ff:ff:ff:ff:ff:ff
    inet XXX.XXX.XXX.XXX/27 brd XXX.XXX.XXX.XXX scope global eth0

Solution 1

Looks like there are a solution since nginx >= 1.5.8:

 resolver 8.8.8.8 ipv6=off;

You can find more here: http://nginx.org/en/docs/http/ngx_http_core_module.html

Solution 2

Nginx is doing DNS resolution at startup by default.

Include (an empty) variable in the hostname to force Nginx to do resolution at runtime with the specified resolver directive.

location / {
    resolver 1.1.1.1 ipv6=off valid=30s;
    set $empty "";
    proxy_pass https://example.com$empty;
}

Solution 3

Using the resolver did not work for me when using proxy_pass to a https url. I had to modify the sysctl.

1. Add the following lines in /etc/sysctl.conf.
   net.ipv6.conf.all.disable_ipv6 = 1 net.ipv6.conf.default.disable_ipv6 = 1 net.ipv6.conf.lo.disable_ipv6 = 1 net.ipv6.conf.eth0.disable_ipv6 = 1 net.ipv6.conf.eth1.disable_ipv6 = 1 net.ipv6.conf.eth2.disable_ipv6 = 1 net.ipv6.conf.eth3.disable_ipv6 = 1
2. Reboot system with sysctl -p.
3. Reboot nginx with sudo nginx -s reload.

Solution 4

I had the same issue with fastcgi_pass to localhost. In this special case (i.e., if you are only forwarding to localhost), you can fix this by using

fastcgi_pass 127.0.0.1:9000;

instead of

fastcgi_pass localhost:9000;

View more solutions

Comments

[Anton]

Flup, net.ipv6.bindv6only=0 doesn't help

[Michael Hampton]

That definitely looks wrong. You should have at least IPv6 link-local addresses. What changes have you or your provider made to this server's configuration?

[Anton]

We've added only net.ipv6.conf.all.disable_ipv6=1. Our provider doesn't assign IPv6 using DHCP, so it must be configured manually.

[Dmitry Polushkin]

Is there are anyway to solve that specifically with nginx config, without changing global system settings?

[Michael Hampton]

You should not use the disable_ipv6 sysctl, and you should configure IPv6 if you have it available to you.

[Pascal Pixel Rigaux]

NB: by default nginx handles IPv4 & IPv6 as different "upstream"s. It will use round robin to choose which to use (cf function "ngx_http_upstream_init_round_robin" which uses "ngx_inet_resolve_host")

[GaryBishop]

Solved my problem proxying flickr.

[GaryBishop]

It fixed my problem but now the error log is showing complaints (99: Cannot assign requested address) even though the request apparently succeeds. I'd just like to figure out how to prevent it from trying the ipv6 address.

[GaryBishop]

Adding the ipv6=off to the resolver in nginx along with this fixes it and eliminates the error messages.

[phiresky]

this works but you also need to add $request_uri because it is only added by default when there is no other variable present

[Pascal Pixel Rigaux]

As explained by Pedro, this works for dynamic DNS resolution (when proxy_pass host is dynamic): you must use a variable in the host: http://${empty}example.com

[GaryBishop]

It won't allow me to set the variable empty. I read elsewhere that custom variables aren't allowed. What are you doing to allow the set $empty ""?

[piotrekkr]

@phiresky $request_uri is automatically added when there is no /uri/ present in proxy_pass url. See here nginx.org/en/docs/http/ngx_http_proxy_module.html#proxy_pass

[Bennet G.]

This completely disables ipv6 on the OS. Which can be done if you don't want any ipv6 traffic in our out, but might overshoot on not-resolving ipv6 to your proxy-pass.