Module ngx_stream_core_module

The ngx_stream_core_module module is available since version 1.9.0. This module is not built by default, it should be enabled with the --with-stream configuration parameter.

worker_processes auto;

error_log /var/log/nginx/error.log info;

events {
    worker_connections  1024;
}

stream {
    upstream backend {
        hash $remote_addr consistent;

        server backend1.example.com:12345 weight=5;
        server 127.0.0.1:12345            max_fails=3 fail_timeout=30s;
        server unix:/tmp/backend3;
    }

    upstream dns {
       server 192.168.0.1:53535;
       server dns.example.com:53;
    }

    server {
        listen 12345;
        proxy_connect_timeout 1s;
        proxy_timeout 3s;
        proxy_pass backend;
    }

    server {
        listen 127.0.0.1:53 udp reuseport;
        proxy_timeout 20s;
        proxy_pass dns;
    }

    server {
        listen [::1]:12345;
        proxy_pass unix:/tmp/stream.socket;
    }
}

Sets the address and port for the socket on which the server will accept connections. It is possible to specify just the port. The address can also be a hostname, for example:

listen 127.0.0.1:12345;
listen *:12345;
listen 12345;     # same as *:12345
listen localhost:12345;

IPv6 addresses are specified in square brackets:

listen [::1]:12345;
listen [::]:12345;

UNIX-domain sockets are specified with the “unix:” prefix:

listen unix:/var/run/nginx.sock;

Port ranges (1.15.10) are specified with the first and last port separated by a hyphen:

listen 127.0.0.1:12345-12399;
listen 12345-12399;

The default_server parameter, if present, will cause the server to become the default server for the specified address:port pair (1.25.5). If none of the directives have the default_server parameter then the first server with the address:port pair will be the default server for this pair.

The ssl parameter allows specifying that all connections accepted on this port should work in SSL mode.

The udp parameter configures a listening socket for working with datagrams (1.9.13). In order to handle packets from the same address and port in the same session, the reuseport parameter should also be specified.

The proxy_protocol parameter (1.11.4) allows specifying that all connections accepted on this port should use the PROXY protocol.

The PROXY protocol version 2 is supported since version 1.13.11.

The listen directive can have several additional parameters specific to socket-related system calls. These parameters can be specified in any listen directive, but only once for a given address:port pair.

setfib=number

this parameter (1.25.5) sets the associated routing table, FIB (the SO_SETFIB option) for the listening socket. This currently works only on FreeBSD.

fastopen=number

enables “TCP Fast Open” for the listening socket (1.21.0) and limits the maximum length for the queue of connections that have not yet completed the three-way handshake.

Do not enable this feature unless the server can handle receiving the same SYN packet with data more than once.

backlog=number

sets the backlog parameter in the listen() call that limits the maximum length for the queue of pending connections (1.9.2). By default, backlog is set to -1 on FreeBSD, DragonFly BSD, and macOS, and to 511 on other platforms.

rcvbuf=size

sets the receive buffer size (the SO_RCVBUF option) for the listening socket (1.11.13).

sndbuf=size

sets the send buffer size (the SO_SNDBUF option) for the listening socket (1.11.13).

accept_filter=filter

sets the name of accept filter (the SO_ACCEPTFILTER option) for the listening socket that filters incoming connections before passing them to accept() (1.25.5). This works only on FreeBSD and NetBSD 5.0+. Possible values are dataready and httpready.

deferred

instructs to use a deferred accept() (the TCP_DEFER_ACCEPT socket option) on Linux (1.25.5).

bind

this parameter instructs to make a separate bind() call for a given address:port pair. The fact is that if there are several listen directives with the same port but different addresses, and one of the listen directives listens on all addresses for the given port (*:port), nginx will bind() only to *:port. It should be noted that the getsockname() system call will be made in this case to determine the address that accepted the connection. If the setfib, fastopen, backlog, rcvbuf, sndbuf, accept_filter, deferred, ipv6only, reuseport, or so_keepalive parameters are used then for a given address:port pair a separate bind() call will always be made.

ipv6only=on|off

this parameter determines (via the IPV6_V6ONLY socket option) whether an IPv6 socket listening on a wildcard address [::] will accept only IPv6 connections or both IPv6 and IPv4 connections. This parameter is turned on by default. It can only be set once on start.

reuseport

this parameter (1.9.1) instructs to create an individual listening socket for each worker process (using the SO_REUSEPORT socket option on Linux 3.9+ and DragonFly BSD, or SO_REUSEPORT_LB on FreeBSD 12+), allowing a kernel to distribute incoming connections between worker processes. This currently works only on Linux 3.9+, DragonFly BSD, and FreeBSD 12+ (1.15.1).

Inappropriate use of this option may have its security implications.

so_keepalive=on|off|[keepidle]:[keepintvl]:[keepcnt]

this parameter configures the “TCP keepalive” behavior for the listening socket. If this parameter is omitted then the operating system’s settings will be in effect for the socket. If it is set to the value “on”, the SO_KEEPALIVE option is turned on for the socket. If it is set to the value “off”, the SO_KEEPALIVE option is turned off for the socket. Some operating systems support setting of TCP keepalive parameters on a per-socket basis using the TCP_KEEPIDLE, TCP_KEEPINTVL, and TCP_KEEPCNT socket options. On such systems (currently, Linux 2.4+, NetBSD 5+, and FreeBSD 9.0-STABLE), they can be configured using the keepidle, keepintvl, and keepcnt parameters. One or two parameters may be omitted, in which case the system default setting for the corresponding socket option will be in effect. For example,

so_keepalive=30m::10

will set the idle timeout (TCP_KEEPIDLE) to 30 minutes, leave the probe interval (TCP_KEEPINTVL) at its system default, and set the probes count (TCP_KEEPCNT) to 10 probes.

Before version 1.25.5, different servers must listen on different address:port pairs.

preread_buffer_size 16k;

Specifies a size of the preread buffer.

preread_timeout 30s;

Specifies a timeout of the preread phase.

proxy_protocol_timeout 30s;

Specifies a timeout for reading the PROXY protocol header to complete. If no entire header is transmitted within this time, the connection is closed.

Configures name servers used to resolve names of upstream servers into addresses, for example:

resolver 127.0.0.1 [::1]:5353;

The address can be specified as a domain name or IP address, with an optional port. If port is not specified, the port 53 is used. Name servers are queried in a round-robin fashion.

By default, nginx will look up both IPv4 and IPv6 addresses while resolving. If looking up of IPv4 or IPv6 addresses is not desired, the ipv4=off (1.23.1) or the ipv6=off parameter can be specified.

By default, nginx caches answers using the TTL value of a response. The optional valid parameter allows overriding it:

resolver 127.0.0.1 [::1]:5353 valid=30s;

To prevent DNS spoofing, it is recommended configuring DNS servers in a properly secured trusted local network.

The optional status_zone parameter (1.17.1) enables collection of DNS server statistics of requests and responses in the specified zone. The parameter is available as part of our commercial subscription.

Before version 1.11.3, this directive was available as part of our commercial subscription.

resolver_timeout 30s;

Sets a timeout for name resolution, for example:

resolver_timeout 5s;

Before version 1.11.3, this directive was available as part of our commercial subscription.

Sets the configuration for a virtual server. There is no clear separation between IP-based (based on the IP address) and name-based (based on the TLS Server Name Indication extension (SNI, RFC 6066)) (1.25.5) virtual servers. Instead, the listen directives describe all addresses and ports that should accept connections for the server, and the server_name directive lists all server names.

server_name "";

Sets names of a virtual server, for example:

server {
    server_name example.com www.example.com;
}

The first name becomes the primary server name.

Server names can include an asterisk (“*”) replacing the first or last part of a name:

server {
    server_name example.com *.example.com www.example.*;
}

Such names are called wildcard names.

The first two of the names mentioned above can be combined in one:

server {
    server_name .example.com;
}

It is also possible to use regular expressions in server names, preceding the name with a tilde (“~”):

server {
    server_name www.example.com ~^www\d+\.example\.com$;
}

Regular expressions can contain captures that can later be used in other directives:

server {
    server_name ~^(www\.)?(.+)$;

    proxy_pass www.$2:12345;
}

Named captures in regular expressions create variables that can later be used in other directives:

server {
    server_name ~^(www\.)?(?<domain>.+)$;

    proxy_pass www.$domain:12345;
}

If the directive’s parameter is set to “$hostname”, the machine’s hostname is inserted.

During searching for a virtual server by name, if the name matches more than one of the specified variants, (e.g. both a wildcard name and regular expression match), the first matching variant will be chosen, in the following order of priority:

1. the exact name
2. the longest wildcard name starting with an asterisk, e.g. “*.example.com”
3. the longest wildcard name ending with an asterisk, e.g. “mail.*”
4. the first matching regular expression (in order of appearance in the configuration file)

server_names_hash_bucket_size 32|64|128;

Sets the bucket size for the server names hash tables. The default value depends on the size of the processor’s cache line. The details of setting up hash tables are provided in a separate document.

server_names_hash_max_size 512;

Sets the maximum size of the server names hash tables. The details of setting up hash tables are provided in a separate document.

Provides the configuration file context in which the stream server directives are specified.

tcp_nodelay on;

Enables or disables the use of the TCP_NODELAY option. The option is enabled for both client and proxied server connections.

variables_hash_bucket_size 64;

Sets the bucket size for the variables hash table. The details of setting up hash tables are provided in a separate document.

variables_hash_max_size 1024;

Sets the maximum size of the variables hash table. The details of setting up hash tables are provided in a separate document.

The ngx_stream_core_module module supports variables since 1.11.2.

$binary_remote_addr

client address in a binary form, value’s length is always 4 bytes for IPv4 addresses or 16 bytes for IPv6 addresses

$bytes_received

number of bytes received from a client (1.11.4)

$bytes_sent

number of bytes sent to a client

$connection

connection serial number

$hostname

host name

$msec

current time in seconds with the milliseconds resolution

$nginx_version

nginx version

$pid

PID of the worker process

$protocol

protocol used to communicate with the client: TCP or UDP (1.11.4)

$proxy_protocol_addr

client address from the PROXY protocol header (1.11.4)

The PROXY protocol must be previously enabled by setting the proxy_protocol parameter in the listen directive.

$proxy_protocol_port

client port from the PROXY protocol header (1.11.4)

The PROXY protocol must be previously enabled by setting the proxy_protocol parameter in the listen directive.

$proxy_protocol_server_addr

server address from the PROXY protocol header (1.17.6)

The PROXY protocol must be previously enabled by setting the proxy_protocol parameter in the listen directive.

$proxy_protocol_server_port

server port from the PROXY protocol header (1.17.6)

The PROXY protocol must be previously enabled by setting the proxy_protocol parameter in the listen directive.

$proxy_protocol_tlv_name

TLV from the PROXY Protocol header (1.23.2). The name can be a TLV type name or its numeric value. In the latter case, the value is hexadecimal and should be prefixed with 0x:

$proxy_protocol_tlv_alpn
$proxy_protocol_tlv_0x01

SSL TLVs can also be accessed by TLV type name or its numeric value, both prefixed by ssl_:

$proxy_protocol_tlv_ssl_version
$proxy_protocol_tlv_ssl_0x21

The following TLV type names are supported:

• alpn (0x01) - upper layer protocol used over the connection
• authority (0x02) - host name value passed by the client
• unique_id (0x05) - unique connection id
• netns (0x30) - name of the namespace
• ssl (0x20) - binary SSL TLV structure

The following SSL TLV type names are supported:

• ssl_version (0x21) - SSL version used in client connection
• ssl_cn (0x22) - SSL certificate Common Name
• ssl_cipher (0x23) - name of the used cipher
• ssl_sig_alg (0x24) - algorithm used to sign the certificate
• ssl_key_alg (0x25) - public-key algorithm

Also, the following special SSL TLV type name is supported:

• ssl_verify - client SSL certificate verification result, zero if the client presented a certificate and it was successfully verified, and non-zero otherwise

The PROXY protocol must be previously enabled by setting the proxy_protocol parameter in the listen directive.

$remote_addr

client address

$remote_port

client port

$server_addr

an address of the server which accepted a connection

Computing a value of this variable usually requires one system call. To avoid a system call, the listen directives must specify addresses and use the bind parameter.

$server_port

port of the server which accepted a connection

$session_time

session duration in seconds with a milliseconds resolution (1.11.4);

$status

session status (1.11.4), can be one of the following:

200

session completed successfully

400

client data could not be parsed, for example, the PROXY protocol header

403

access forbidden, for example, when access is limited for certain client addresses

500

internal server error

502

bad gateway, for example, if an upstream server could not be selected or reached.

503

service unavailable, for example, when access is limited by the number of connections

$time_iso8601

local time in the ISO 8601 standard format

$time_local

local time in the Common Log Format