nghttp2/doc/nghttpx.1

2770 lines
91 KiB
Groff

.\" Man page generated from reStructuredText.
.
.
.nr rst2man-indent-level 0
.
.de1 rstReportMargin
\\$1 \\n[an-margin]
level \\n[rst2man-indent-level]
level margin: \\n[rst2man-indent\\n[rst2man-indent-level]]
-
\\n[rst2man-indent0]
\\n[rst2man-indent1]
\\n[rst2man-indent2]
..
.de1 INDENT
.\" .rstReportMargin pre:
. RS \\$1
. nr rst2man-indent\\n[rst2man-indent-level] \\n[an-margin]
. nr rst2man-indent-level +1
.\" .rstReportMargin post:
..
.de UNINDENT
. RE
.\" indent \\n[an-margin]
.\" old: \\n[rst2man-indent\\n[rst2man-indent-level]]
.nr rst2man-indent-level -1
.\" new: \\n[rst2man-indent\\n[rst2man-indent-level]]
.in \\n[rst2man-indent\\n[rst2man-indent-level]]u
..
.TH "NGHTTPX" "1" "Sep 21, 2022" "1.50.0" "nghttp2"
.SH NAME
nghttpx \- HTTP/2 proxy
.SH SYNOPSIS
.sp
\fBnghttpx\fP [OPTIONS]... [<PRIVATE_KEY> <CERT>]
.SH DESCRIPTION
.sp
A reverse proxy for HTTP/3, HTTP/2, and HTTP/1.
.INDENT 0.0
.TP
.B <PRIVATE_KEY>
Set path to server\(aqs private key. Required unless
"no\-tls" parameter is used in \fI\%\-\-frontend\fP option.
.UNINDENT
.INDENT 0.0
.TP
.B <CERT>
Set path to server\(aqs certificate. Required unless
"no\-tls" parameter is used in \fI\%\-\-frontend\fP option. To
make OCSP stapling work, this must be an absolute path.
.UNINDENT
.SH OPTIONS
.sp
The options are categorized into several groups.
.SS Connections
.INDENT 0.0
.TP
.B \-b, \-\-backend=(<HOST>,<PORT>|unix:<PATH>)[;[<PATTERN>[:...]][[;<PARAM>]...]
Set backend host and port. The multiple backend
addresses are accepted by repeating this option. UNIX
domain socket can be specified by prefixing path name
with "unix:" (e.g., unix:/var/run/backend.sock).
.sp
Optionally, if <PATTERN>s are given, the backend address
is only used if request matches the pattern. The
pattern matching is closely designed to ServeMux in
net/http package of Go programming language. <PATTERN>
consists of path, host + path or just host. The path
must start with "\fI/\fP". If it ends with "\fI/\fP", it matches
all request path in its subtree. To deal with the
request to the directory without trailing slash, the
path which ends with "\fI/\fP" also matches the request path
which only lacks trailing \(aq\fI/\fP\(aq (e.g., path "\fI/foo/\fP"
matches request path "\fI/foo\fP"). If it does not end with
"\fI/\fP", it performs exact match against the request path.
If host is given, it performs a match against the
request host. For a request received on the frontend
listener with "sni\-fwd" parameter enabled, SNI host is
used instead of a request host. If host alone is given,
"\fI/\fP" is appended to it, so that it matches all request
paths under the host (e.g., specifying "nghttp2.org"
equals to "nghttp2.org/"). CONNECT method is treated
specially. It does not have path, and we don\(aqt allow
empty path. To workaround this, we assume that CONNECT
method has "\fI/\fP" as path.
.sp
Patterns with host take precedence over patterns with
just path. Then, longer patterns take precedence over
shorter ones.
.sp
Host can include "*" in the left most position to
indicate wildcard match (only suffix match is done).
The "*" must match at least one character. For example,
host pattern "*.nghttp2.org" matches against
"www.nghttp2.org" and "git.ngttp2.org", but does not
match against "nghttp2.org". The exact hosts match
takes precedence over the wildcard hosts match.
.sp
If path part ends with "*", it is treated as wildcard
path. The wildcard path behaves differently from the
normal path. For normal path, match is made around the
boundary of path component separator,"\fI/\fP". On the other
hand, the wildcard path does not take into account the
path component separator. All paths which include the
wildcard path without last "*" as prefix, and are
strictly longer than wildcard path without last "*" are
matched. "*" must match at least one character. For
example, the pattern "\fI/foo*\fP" matches "\fI/foo/\fP" and
"\fI/foobar\fP". But it does not match "\fI/foo\fP", or "\fI/fo\fP".
.sp
If <PATTERN> is omitted or empty string, "\fI/\fP" is used as
pattern, which matches all request paths (catch\-all
pattern). The catch\-all backend must be given.
.sp
When doing a match, nghttpx made some normalization to
pattern, request host and path. For host part, they are
converted to lower case. For path part, percent\-encoded
unreserved characters defined in RFC 3986 are decoded,
and any dot\-segments (".." and ".") are resolved and
removed.
.sp
For example, \fI\%\-b\fP\(aq127.0.0.1,8080;nghttp2.org/httpbin/\(aq
matches the request host "nghttp2.org" and the request
path "\fI/httpbin/get\fP", but does not match the request host
"nghttp2.org" and the request path "\fI/index.html\fP".
.sp
The multiple <PATTERN>s can be specified, delimiting
them by ":". Specifying
\fI\%\-b\fP\(aq127.0.0.1,8080;nghttp2.org:www.nghttp2.org\(aq has the
same effect to specify \fI\%\-b\fP\(aq127.0.0.1,8080;nghttp2.org\(aq
and \fI\%\-b\fP\(aq127.0.0.1,8080;www.nghttp2.org\(aq.
.sp
The backend addresses sharing same <PATTERN> are grouped
together forming load balancing group.
.sp
Several parameters <PARAM> are accepted after <PATTERN>.
The parameters are delimited by ";". The available
parameters are: "proto=<PROTO>", "tls",
"sni=<SNI_HOST>", "fall=<N>", "rise=<N>",
"affinity=<METHOD>", "dns", "redirect\-if\-not\-tls",
"upgrade\-scheme", "mruby=<PATH>",
"read\-timeout=<DURATION>", "write\-timeout=<DURATION>",
"group=<GROUP>", "group\-weight=<N>", "weight=<N>", and
"dnf". The parameter consists of keyword, and
optionally followed by "=" and value. For example, the
parameter "proto=h2" consists of the keyword "proto" and
value "h2". The parameter "tls" consists of the keyword
"tls" without value. Each parameter is described as
follows.
.sp
The backend application protocol can be specified using
optional "proto" parameter, and in the form of
"proto=<PROTO>". <PROTO> should be one of the following
list without quotes: "h2", "http/1.1". The default
value of <PROTO> is "http/1.1". Note that usually "h2"
refers to HTTP/2 over TLS. But in this option, it may
mean HTTP/2 over cleartext TCP unless "tls" keyword is
used (see below).
.sp
TLS can be enabled by specifying optional "tls"
parameter. TLS is not enabled by default.
.sp
With "sni=<SNI_HOST>" parameter, it can override the TLS
SNI field value with given <SNI_HOST>. This will
default to the backend <HOST> name
.sp
The feature to detect whether backend is online or
offline can be enabled using optional "fall" and "rise"
parameters. Using "fall=<N>" parameter, if nghttpx
cannot connect to a this backend <N> times in a row,
this backend is assumed to be offline, and it is
excluded from load balancing. If <N> is 0, this backend
never be excluded from load balancing whatever times
nghttpx cannot connect to it, and this is the default.
There is also "rise=<N>" parameter. After backend was
excluded from load balancing group, nghttpx periodically
attempts to make a connection to the failed backend, and
if the connection is made successfully <N> times in a
row, the backend is assumed to be online, and it is now
eligible for load balancing target. If <N> is 0, a
backend is permanently offline, once it goes in that
state, and this is the default behaviour.
.sp
The session affinity is enabled using
"affinity=<METHOD>" parameter. If "ip" is given in
<METHOD>, client IP based session affinity is enabled.
If "cookie" is given in <METHOD>, cookie based session
affinity is enabled. If "none" is given in <METHOD>,
session affinity is disabled, and this is the default.
The session affinity is enabled per <PATTERN>. If at
least one backend has "affinity" parameter, and its
<METHOD> is not "none", session affinity is enabled for
all backend servers sharing the same <PATTERN>. It is
advised to set "affinity" parameter to all backend
explicitly if session affinity is desired. The session
affinity may break if one of the backend gets
unreachable, or backend settings are reloaded or
replaced by API.
.sp
If "affinity=cookie" is used, the additional
configuration is required.
"affinity\-cookie\-name=<NAME>" must be used to specify a
name of cookie to use. Optionally,
"affinity\-cookie\-path=<PATH>" can be used to specify a
path which cookie is applied. The optional
"affinity\-cookie\-secure=<SECURE>" controls the Secure
attribute of a cookie. The default value is "auto", and
the Secure attribute is determined by a request scheme.
If a request scheme is "https", then Secure attribute is
set. Otherwise, it is not set. If <SECURE> is "yes",
the Secure attribute is always set. If <SECURE> is
"no", the Secure attribute is always omitted.
"affinity\-cookie\-stickiness=<STICKINESS>" controls
stickiness of this affinity. If <STICKINESS> is
"loose", removing or adding a backend server might break
the affinity and the request might be forwarded to a
different backend server. If <STICKINESS> is "strict",
removing the designated backend server breaks affinity,
but adding new backend server does not cause breakage.
If the designated backend server becomes unavailable,
new backend server is chosen as if the request does not
have an affinity cookie. <STICKINESS> defaults to
"loose".
.sp
By default, name resolution of backend host name is done
at start up, or reloading configuration. If "dns"
parameter is given, name resolution takes place
dynamically. This is useful if backend address changes
frequently. If "dns" is given, name resolution of
backend host name at start up, or reloading
configuration is skipped.
.sp
If "redirect\-if\-not\-tls" parameter is used, the matched
backend requires that frontend connection is TLS
encrypted. If it isn\(aqt, nghttpx responds to the request
with 308 status code, and https URI the client should
use instead is included in Location header field. The
port number in redirect URI is 443 by default, and can
be changed using \fI\%\-\-redirect\-https\-port\fP option. If at
least one backend has "redirect\-if\-not\-tls" parameter,
this feature is enabled for all backend servers sharing
the same <PATTERN>. It is advised to set
"redirect\-if\-no\-tls" parameter to all backends
explicitly if this feature is desired.
.sp
If "upgrade\-scheme" parameter is used along with "tls"
parameter, HTTP/2 :scheme pseudo header field is changed
to "https" from "http" when forwarding a request to this
particular backend. This is a workaround for a backend
server which requires "https" :scheme pseudo header
field on TLS encrypted connection.
.sp
"mruby=<PATH>" parameter specifies a path to mruby
script file which is invoked when this pattern is
matched. All backends which share the same pattern must
have the same mruby path.
.sp
"read\-timeout=<DURATION>" and "write\-timeout=<DURATION>"
parameters specify the read and write timeout of the
backend connection when this pattern is matched. All
backends which share the same pattern must have the same
timeouts. If these timeouts are entirely omitted for a
pattern, \fI\%\-\-backend\-read\-timeout\fP and
\fI\%\-\-backend\-write\-timeout\fP are used.
.sp
"group=<GROUP>" parameter specifies the name of group
this backend address belongs to. By default, it belongs
to the unnamed default group. The name of group is
unique per pattern. "group\-weight=<N>" parameter
specifies the weight of the group. The higher weight
gets more frequently selected by the load balancing
algorithm. <N> must be [1, 256] inclusive. The weight
8 has 4 times more weight than 2. <N> must be the same
for all addresses which share the same <GROUP>. If
"group\-weight" is omitted in an address, but the other
address which belongs to the same group specifies
"group\-weight", its weight is used. If no
"group\-weight" is specified for all addresses, the
weight of a group becomes 1. "group" and "group\-weight"
are ignored if session affinity is enabled.
.sp
"weight=<N>" parameter specifies the weight of the
backend address inside a group which this address
belongs to. The higher weight gets more frequently
selected by the load balancing algorithm. <N> must be
[1, 256] inclusive. The weight 8 has 4 times more
weight than weight 2. If this parameter is omitted,
weight becomes 1. "weight" is ignored if session
affinity is enabled.
.sp
If "dnf" parameter is specified, an incoming request is
not forwarded to a backend and just consumed along with
the request body (actually a backend server never be
contacted). It is expected that the HTTP response is
generated by mruby script (see "mruby=<PATH>" parameter
above). "dnf" is an abbreviation of "do not forward".
.sp
Since ";" and ":" are used as delimiter, <PATTERN> must
not contain these characters. In order to include ":"
in <PATTERN>, one has to specify "%3A" (which is
percent\-encoded from of ":") instead. Since ";" has
special meaning in shell, the option value must be
quoted.
.sp
Default: \fB127.0.0.1,80\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-f, \-\-frontend=(<HOST>,<PORT>|unix:<PATH>)[[;<PARAM>]...]
Set frontend host and port. If <HOST> is \(aq*\(aq, it
assumes all addresses including both IPv4 and IPv6.
UNIX domain socket can be specified by prefixing path
name with "unix:" (e.g., unix:/var/run/nghttpx.sock).
This option can be used multiple times to listen to
multiple addresses.
.sp
This option can take 0 or more parameters, which are
described below. Note that "api" and "healthmon"
parameters are mutually exclusive.
.sp
Optionally, TLS can be disabled by specifying "no\-tls"
parameter. TLS is enabled by default.
.sp
If "sni\-fwd" parameter is used, when performing a match
to select a backend server, SNI host name received from
the client is used instead of the request host. See
\fI\%\-\-backend\fP option about the pattern match.
.sp
To make this frontend as API endpoint, specify "api"
parameter. This is disabled by default. It is
important to limit the access to the API frontend.
Otherwise, someone may change the backend server, and
break your services, or expose confidential information
to the outside the world.
.sp
To make this frontend as health monitor endpoint,
specify "healthmon" parameter. This is disabled by
default. Any requests which come through this address
are replied with 200 HTTP status, without no body.
.sp
To accept PROXY protocol version 1 and 2 on frontend
connection, specify "proxyproto" parameter. This is
disabled by default.
.sp
To receive HTTP/3 (QUIC) traffic, specify "quic"
parameter. It makes nghttpx listen on UDP port rather
than TCP port. UNIX domain socket, "api", and
"healthmon" parameters cannot be used with "quic"
parameter.
.sp
Default: \fB*,3000\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backlog=<N>
Set listen backlog size.
.sp
Default: \fB65536\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-address\-family=(auto|IPv4|IPv6)
Specify address family of backend connections. If
"auto" is given, both IPv4 and IPv6 are considered. If
"IPv4" is given, only IPv4 address is considered. If
"IPv6" is given, only IPv6 address is considered.
.sp
Default: \fBauto\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-http\-proxy\-uri=<URI>
Specify proxy URI in the form
\fI\%http:/\fP/[<USER>:<PASS>@]<PROXY>:<PORT>. If a proxy
requires authentication, specify <USER> and <PASS>.
Note that they must be properly percent\-encoded. This
proxy is used when the backend connection is HTTP/2.
First, make a CONNECT request to the proxy and it
connects to the backend on behalf of nghttpx. This
forms tunnel. After that, nghttpx performs SSL/TLS
handshake with the downstream through the tunnel. The
timeouts when connecting and making CONNECT request can
be specified by \fI\%\-\-backend\-read\-timeout\fP and
\fI\%\-\-backend\-write\-timeout\fP options.
.UNINDENT
.SS Performance
.INDENT 0.0
.TP
.B \-n, \-\-workers=<N>
Set the number of worker threads.
.sp
Default: \fB1\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-single\-thread
Run everything in one thread inside the worker process.
This feature is provided for better debugging
experience, or for the platforms which lack thread
support. If threading is disabled, this option is
always enabled.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-read\-rate=<SIZE>
Set maximum average read rate on frontend connection.
Setting 0 to this option means read rate is unlimited.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-read\-burst=<SIZE>
Set maximum read burst size on frontend connection.
Setting 0 to this option means read burst size is
unlimited.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-write\-rate=<SIZE>
Set maximum average write rate on frontend connection.
Setting 0 to this option means write rate is unlimited.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-write\-burst=<SIZE>
Set maximum write burst size on frontend connection.
Setting 0 to this option means write burst size is
unlimited.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-worker\-read\-rate=<SIZE>
Set maximum average read rate on frontend connection per
worker. Setting 0 to this option means read rate is
unlimited. Not implemented yet.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-worker\-read\-burst=<SIZE>
Set maximum read burst size on frontend connection per
worker. Setting 0 to this option means read burst size
is unlimited. Not implemented yet.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-worker\-write\-rate=<SIZE>
Set maximum average write rate on frontend connection
per worker. Setting 0 to this option means write rate
is unlimited. Not implemented yet.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-worker\-write\-burst=<SIZE>
Set maximum write burst size on frontend connection per
worker. Setting 0 to this option means write burst size
is unlimited. Not implemented yet.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-worker\-frontend\-connections=<N>
Set maximum number of simultaneous connections frontend
accepts. Setting 0 means unlimited.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-connections\-per\-host=<N>
Set maximum number of backend concurrent connections
(and/or streams in case of HTTP/2) per origin host.
This option is meaningful when \fI\%\-\-http2\-proxy\fP option is
used. The origin host is determined by authority
portion of request URI (or :authority header field for
HTTP/2). To limit the number of connections per
frontend for default mode, use
\fI\%\-\-backend\-connections\-per\-frontend\fP\&.
.sp
Default: \fB8\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-connections\-per\-frontend=<N>
Set maximum number of backend concurrent connections
(and/or streams in case of HTTP/2) per frontend. This
option is only used for default mode. 0 means
unlimited. To limit the number of connections per host
with \fI\%\-\-http2\-proxy\fP option, use
\fI\%\-\-backend\-connections\-per\-host\fP\&.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-rlimit\-nofile=<N>
Set maximum number of open files (RLIMIT_NOFILE) to <N>.
If 0 is given, nghttpx does not set the limit.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-rlimit\-memlock=<N>
Set maximum number of bytes of memory that may be locked
into RAM. If 0 is given, nghttpx does not set the
limit.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-request\-buffer=<SIZE>
Set buffer size used to store backend request.
.sp
Default: \fB16K\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-response\-buffer=<SIZE>
Set buffer size used to store backend response.
.sp
Default: \fB128K\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-fastopen=<N>
Enables "TCP Fast Open" for the listening socket and
limits the maximum length for the queue of connections
that have not yet completed the three\-way handshake. If
value is 0 then fast open is disabled.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-no\-kqueue
Don\(aqt use kqueue. This option is only applicable for
the platforms which have kqueue. For other platforms,
this option will be simply ignored.
.UNINDENT
.SS Timeout
.INDENT 0.0
.TP
.B \-\-frontend\-http2\-read\-timeout=<DURATION>
Specify read timeout for HTTP/2 frontend connection.
.sp
Default: \fB3m\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http3\-read\-timeout=<DURATION>
Specify read timeout for HTTP/3 frontend connection.
.sp
Default: \fB3m\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-read\-timeout=<DURATION>
Specify read timeout for HTTP/1.1 frontend connection.
.sp
Default: \fB1m\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-write\-timeout=<DURATION>
Specify write timeout for all frontend connections.
.sp
Default: \fB30s\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-keep\-alive\-timeout=<DURATION>
Specify keep\-alive timeout for frontend HTTP/1
connection.
.sp
Default: \fB1m\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-stream\-read\-timeout=<DURATION>
Specify read timeout for HTTP/2 streams. 0 means no
timeout.
.sp
Default: \fB0\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-stream\-write\-timeout=<DURATION>
Specify write timeout for HTTP/2 streams. 0 means no
timeout.
.sp
Default: \fB1m\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-read\-timeout=<DURATION>
Specify read timeout for backend connection.
.sp
Default: \fB1m\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-write\-timeout=<DURATION>
Specify write timeout for backend connection.
.sp
Default: \fB30s\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-connect\-timeout=<DURATION>
Specify timeout before establishing TCP connection to
backend.
.sp
Default: \fB30s\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-keep\-alive\-timeout=<DURATION>
Specify keep\-alive timeout for backend HTTP/1
connection.
.sp
Default: \fB2s\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-listener\-disable\-timeout=<DURATION>
After accepting connection failed, connection listener
is disabled for a given amount of time. Specifying 0
disables this feature.
.sp
Default: \fB30s\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http2\-setting\-timeout=<DURATION>
Specify timeout before SETTINGS ACK is received from
client.
.sp
Default: \fB10s\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-http2\-settings\-timeout=<DURATION>
Specify timeout before SETTINGS ACK is received from
backend server.
.sp
Default: \fB10s\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-max\-backoff=<DURATION>
Specify maximum backoff interval. This is used when
doing health check against offline backend (see "fail"
parameter in \fI\%\-\-backend\fP option). It is also used to
limit the maximum interval to temporarily disable
backend when nghttpx failed to connect to it. These
intervals are calculated using exponential backoff, and
consecutive failed attempts increase the interval. This
option caps its maximum value.
.sp
Default: \fB2m\fP
.UNINDENT
.SS SSL/TLS
.INDENT 0.0
.TP
.B \-\-ciphers=<SUITE>
Set allowed cipher list for frontend connection. The
format of the string is described in OpenSSL ciphers(1).
This option sets cipher suites for TLSv1.2 or earlier.
Use \fI\%\-\-tls13\-ciphers\fP for TLSv1.3.
.sp
Default: \fBECDHE\-ECDSA\-AES128\-GCM\-SHA256:ECDHE\-RSA\-AES128\-GCM\-SHA256:ECDHE\-ECDSA\-AES256\-GCM\-SHA384:ECDHE\-RSA\-AES256\-GCM\-SHA384:ECDHE\-ECDSA\-CHACHA20\-POLY1305:ECDHE\-RSA\-CHACHA20\-POLY1305:DHE\-RSA\-AES128\-GCM\-SHA256:DHE\-RSA\-AES256\-GCM\-SHA384\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls13\-ciphers=<SUITE>
Set allowed cipher list for frontend connection. The
format of the string is described in OpenSSL ciphers(1).
This option sets cipher suites for TLSv1.3. Use
\fI\%\-\-ciphers\fP for TLSv1.2 or earlier.
.sp
Default: \fBTLS_AES_128_GCM_SHA256:TLS_AES_256_GCM_SHA384:TLS_CHACHA20_POLY1305_SHA256\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-client\-ciphers=<SUITE>
Set allowed cipher list for backend connection. The
format of the string is described in OpenSSL ciphers(1).
This option sets cipher suites for TLSv1.2 or earlier.
Use \fI\%\-\-tls13\-client\-ciphers\fP for TLSv1.3.
.sp
Default: \fBECDHE\-ECDSA\-AES128\-GCM\-SHA256:ECDHE\-RSA\-AES128\-GCM\-SHA256:ECDHE\-ECDSA\-AES256\-GCM\-SHA384:ECDHE\-RSA\-AES256\-GCM\-SHA384:ECDHE\-ECDSA\-CHACHA20\-POLY1305:ECDHE\-RSA\-CHACHA20\-POLY1305:DHE\-RSA\-AES128\-GCM\-SHA256:DHE\-RSA\-AES256\-GCM\-SHA384\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls13\-client\-ciphers=<SUITE>
Set allowed cipher list for backend connection. The
format of the string is described in OpenSSL ciphers(1).
This option sets cipher suites for TLSv1.3. Use
\fI\%\-\-tls13\-client\-ciphers\fP for TLSv1.2 or earlier.
.sp
Default: \fBTLS_AES_128_GCM_SHA256:TLS_AES_256_GCM_SHA384:TLS_CHACHA20_POLY1305_SHA256\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-ecdh\-curves=<LIST>
Set supported curve list for frontend connections.
<LIST> is a colon separated list of curve NID or names
in the preference order. The supported curves depend on
the linked OpenSSL library. This function requires
OpenSSL >= 1.0.2.
.sp
Default: \fBX25519:P\-256:P\-384:P\-521\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-k, \-\-insecure
Don\(aqt verify backend server\(aqs certificate if TLS is
enabled for backend connections.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-cacert=<PATH>
Set path to trusted CA certificate file. It is used in
backend TLS connections to verify peer\(aqs certificate.
It is also used to verify OCSP response from the script
set by \fI\%\-\-fetch\-ocsp\-response\-file\fP\&. The file must be in
PEM format. It can contain multiple certificates. If
the linked OpenSSL is configured to load system wide
certificates, they are loaded at startup regardless of
this option.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-private\-key\-passwd\-file=<PATH>
Path to file that contains password for the server\(aqs
private key. If none is given and the private key is
password protected it\(aqll be requested interactively.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-subcert=<KEYPATH>:<CERTPATH>[[;<PARAM>]...]
Specify additional certificate and private key file.
nghttpx will choose certificates based on the hostname
indicated by client using TLS SNI extension. If nghttpx
is built with OpenSSL >= 1.0.2, the shared elliptic
curves (e.g., P\-256) between client and server are also
taken into consideration. This allows nghttpx to send
ECDSA certificate to modern clients, while sending RSA
based certificate to older clients. This option can be
used multiple times. To make OCSP stapling work,
<CERTPATH> must be absolute path.
.sp
Additional parameter can be specified in <PARAM>. The
available <PARAM> is "sct\-dir=<DIR>".
.sp
"sct\-dir=<DIR>" specifies the path to directory which
contains *.sct files for TLS
signed_certificate_timestamp extension (RFC 6962). This
feature requires OpenSSL >= 1.0.2. See also
\fI\%\-\-tls\-sct\-dir\fP option.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-dh\-param\-file=<PATH>
Path to file that contains DH parameters in PEM format.
Without this option, DHE cipher suites are not
available.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-npn\-list=<LIST>
Comma delimited list of ALPN protocol identifier sorted
in the order of preference. That means most desirable
protocol comes first. This is used in both ALPN and
NPN. The parameter must be delimited by a single comma
only and any white spaces are treated as a part of
protocol string.
.sp
Default: \fBh2,h2\-16,h2\-14,http/1.1\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-verify\-client
Require and verify client certificate.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-verify\-client\-cacert=<PATH>
Path to file that contains CA certificates to verify
client certificate. The file must be in PEM format. It
can contain multiple certificates.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-verify\-client\-tolerate\-expired
Accept expired client certificate. Operator should
handle the expired client certificate by some means
(e.g., mruby script). Otherwise, this option might
cause a security risk.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-client\-private\-key\-file=<PATH>
Path to file that contains client private key used in
backend client authentication.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-client\-cert\-file=<PATH>
Path to file that contains client certificate used in
backend client authentication.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-min\-proto\-version=<VER>
Specify minimum SSL/TLS protocol. The name matching is
done in case\-insensitive manner. The versions between
\fI\%\-\-tls\-min\-proto\-version\fP and \fI\%\-\-tls\-max\-proto\-version\fP are
enabled. If the protocol list advertised by client does
not overlap this range, you will receive the error
message "unknown protocol". If a protocol version lower
than TLSv1.2 is specified, make sure that the compatible
ciphers are included in \fI\%\-\-ciphers\fP option. The default
cipher list only includes ciphers compatible with
TLSv1.2 or above. The available versions are:
TLSv1.3, TLSv1.2, TLSv1.1, and TLSv1.0
.sp
Default: \fBTLSv1.2\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-max\-proto\-version=<VER>
Specify maximum SSL/TLS protocol. The name matching is
done in case\-insensitive manner. The versions between
\fI\%\-\-tls\-min\-proto\-version\fP and \fI\%\-\-tls\-max\-proto\-version\fP are
enabled. If the protocol list advertised by client does
not overlap this range, you will receive the error
message "unknown protocol". The available versions are:
TLSv1.3, TLSv1.2, TLSv1.1, and TLSv1.0
.sp
Default: \fBTLSv1.3\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-ticket\-key\-file=<PATH>
Path to file that contains random data to construct TLS
session ticket parameters. If aes\-128\-cbc is given in
\fI\%\-\-tls\-ticket\-key\-cipher\fP, the file must contain exactly
48 bytes. If aes\-256\-cbc is given in
\fI\%\-\-tls\-ticket\-key\-cipher\fP, the file must contain exactly
80 bytes. This options can be used repeatedly to
specify multiple ticket parameters. If several files
are given, only the first key is used to encrypt TLS
session tickets. Other keys are accepted but server
will issue new session ticket with first key. This
allows session key rotation. Please note that key
rotation does not occur automatically. User should
rearrange files or change options values and restart
nghttpx gracefully. If opening or reading given file
fails, all loaded keys are discarded and it is treated
as if none of this option is given. If this option is
not given or an error occurred while opening or reading
a file, key is generated every 1 hour internally and
they are valid for 12 hours. This is recommended if
ticket key sharing between nghttpx instances is not
required.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-ticket\-key\-memcached=<HOST>,<PORT>[;tls]
Specify address of memcached server to get TLS ticket
keys for session resumption. This enables shared TLS
ticket key between multiple nghttpx instances. nghttpx
does not set TLS ticket key to memcached. The external
ticket key generator is required. nghttpx just gets TLS
ticket keys from memcached, and use them, possibly
replacing current set of keys. It is up to extern TLS
ticket key generator to rotate keys frequently. See
"TLS SESSION TICKET RESUMPTION" section in manual page
to know the data format in memcached entry. Optionally,
memcached connection can be encrypted with TLS by
specifying "tls" parameter.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-ticket\-key\-memcached\-address\-family=(auto|IPv4|IPv6)
Specify address family of memcached connections to get
TLS ticket keys. If "auto" is given, both IPv4 and IPv6
are considered. If "IPv4" is given, only IPv4 address
is considered. If "IPv6" is given, only IPv6 address is
considered.
.sp
Default: \fBauto\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-ticket\-key\-memcached\-interval=<DURATION>
Set interval to get TLS ticket keys from memcached.
.sp
Default: \fB10m\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-ticket\-key\-memcached\-max\-retry=<N>
Set maximum number of consecutive retries before
abandoning TLS ticket key retrieval. If this number is
reached, the attempt is considered as failure, and
"failure" count is incremented by 1, which contributed
to the value controlled
\fI\%\-\-tls\-ticket\-key\-memcached\-max\-fail\fP option.
.sp
Default: \fB3\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-ticket\-key\-memcached\-max\-fail=<N>
Set maximum number of consecutive failure before
disabling TLS ticket until next scheduled key retrieval.
.sp
Default: \fB2\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-ticket\-key\-cipher=<CIPHER>
Specify cipher to encrypt TLS session ticket. Specify
either aes\-128\-cbc or aes\-256\-cbc. By default,
aes\-128\-cbc is used.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-ticket\-key\-memcached\-cert\-file=<PATH>
Path to client certificate for memcached connections to
get TLS ticket keys.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-ticket\-key\-memcached\-private\-key\-file=<PATH>
Path to client private key for memcached connections to
get TLS ticket keys.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-fetch\-ocsp\-response\-file=<PATH>
Path to fetch\-ocsp\-response script file. It should be
absolute path.
.sp
Default: \fB/usr/local/share/nghttp2/fetch\-ocsp\-response\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-ocsp\-update\-interval=<DURATION>
Set interval to update OCSP response cache.
.sp
Default: \fB4h\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-ocsp\-startup
Start accepting connections after initial attempts to
get OCSP responses finish. It does not matter some of
the attempts fail. This feature is useful if OCSP
responses must be available before accepting
connections.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-no\-verify\-ocsp
nghttpx does not verify OCSP response.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-no\-ocsp
Disable OCSP stapling.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-session\-cache\-memcached=<HOST>,<PORT>[;tls]
Specify address of memcached server to store session
cache. This enables shared session cache between
multiple nghttpx instances. Optionally, memcached
connection can be encrypted with TLS by specifying "tls"
parameter.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-session\-cache\-memcached\-address\-family=(auto|IPv4|IPv6)
Specify address family of memcached connections to store
session cache. If "auto" is given, both IPv4 and IPv6
are considered. If "IPv4" is given, only IPv4 address
is considered. If "IPv6" is given, only IPv6 address is
considered.
.sp
Default: \fBauto\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-session\-cache\-memcached\-cert\-file=<PATH>
Path to client certificate for memcached connections to
store session cache.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-session\-cache\-memcached\-private\-key\-file=<PATH>
Path to client private key for memcached connections to
store session cache.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-dyn\-rec\-warmup\-threshold=<SIZE>
Specify the threshold size for TLS dynamic record size
behaviour. During a TLS session, after the threshold
number of bytes have been written, the TLS record size
will be increased to the maximum allowed (16K). The max
record size will continue to be used on the active TLS
session. After \fI\%\-\-tls\-dyn\-rec\-idle\-timeout\fP has elapsed,
the record size is reduced to 1300 bytes. Specify 0 to
always use the maximum record size, regardless of idle
period. This behaviour applies to all TLS based
frontends, and TLS HTTP/2 backends.
.sp
Default: \fB1M\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-dyn\-rec\-idle\-timeout=<DURATION>
Specify TLS dynamic record size behaviour timeout. See
\fI\%\-\-tls\-dyn\-rec\-warmup\-threshold\fP for more information.
This behaviour applies to all TLS based frontends, and
TLS HTTP/2 backends.
.sp
Default: \fB1s\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-no\-http2\-cipher\-block\-list
Allow block listed cipher suite on frontend HTTP/2
connection. See
\fI\%https://tools.ietf.org/html/rfc7540#appendix\-A\fP for the
complete HTTP/2 cipher suites block list.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-client\-no\-http2\-cipher\-block\-list
Allow block listed cipher suite on backend HTTP/2
connection. See
\fI\%https://tools.ietf.org/html/rfc7540#appendix\-A\fP for the
complete HTTP/2 cipher suites block list.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-sct\-dir=<DIR>
Specifies the directory where *.sct files exist. All
*.sct files in <DIR> are read, and sent as
extension_data of TLS signed_certificate_timestamp (RFC
6962) to client. These *.sct files are for the
certificate specified in positional command\-line
argument <CERT>, or certificate option in configuration
file. For additional certificates, use \fI\%\-\-subcert\fP
option. This option requires OpenSSL >= 1.0.2.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-psk\-secrets=<PATH>
Read list of PSK identity and secrets from <PATH>. This
is used for frontend connection. The each line of input
file is formatted as <identity>:<hex\-secret>, where
<identity> is PSK identity, and <hex\-secret> is secret
in hex. An empty line, and line which starts with \(aq#\(aq
are skipped. The default enabled cipher list might not
contain any PSK cipher suite. In that case, desired PSK
cipher suites must be enabled using \fI\%\-\-ciphers\fP option.
The desired PSK cipher suite may be block listed by
HTTP/2. To use those cipher suites with HTTP/2,
consider to use \fI\%\-\-no\-http2\-cipher\-block\-list\fP option.
But be aware its implications.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-client\-psk\-secrets=<PATH>
Read PSK identity and secrets from <PATH>. This is used
for backend connection. The each line of input file is
formatted as <identity>:<hex\-secret>, where <identity>
is PSK identity, and <hex\-secret> is secret in hex. An
empty line, and line which starts with \(aq#\(aq are skipped.
The first identity and secret pair encountered is used.
The default enabled cipher list might not contain any
PSK cipher suite. In that case, desired PSK cipher
suites must be enabled using \fI\%\-\-client\-ciphers\fP option.
The desired PSK cipher suite may be block listed by
HTTP/2. To use those cipher suites with HTTP/2,
consider to use \fI\%\-\-client\-no\-http2\-cipher\-block\-list\fP
option. But be aware its implications.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-no\-postpone\-early\-data
By default, except for QUIC connections, nghttpx
postpones forwarding HTTP requests sent in early data,
including those sent in partially in it, until TLS
handshake finishes. If all backend server recognizes
"Early\-Data" header field, using this option makes
nghttpx not postpone forwarding request and get full
potential of 0\-RTT data.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-max\-early\-data=<SIZE>
Sets the maximum amount of 0\-RTT data that server
accepts.
.sp
Default: \fB16K\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-tls\-ktls
Enable ktls. For server, ktls is enable if
\fI\%\-\-tls\-session\-cache\-memcached\fP is not configured.
.UNINDENT
.SS HTTP/2
.INDENT 0.0
.TP
.B \-c, \-\-frontend\-http2\-max\-concurrent\-streams=<N>
Set the maximum number of the concurrent streams in one
frontend HTTP/2 session.
.sp
Default: \fB100\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-http2\-max\-concurrent\-streams=<N>
Set the maximum number of the concurrent streams in one
backend HTTP/2 session. This sets maximum number of
concurrent opened pushed streams. The maximum number of
concurrent requests are set by a remote server.
.sp
Default: \fB100\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http2\-window\-size=<SIZE>
Sets the per\-stream initial window size of HTTP/2
frontend connection.
.sp
Default: \fB65535\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http2\-connection\-window\-size=<SIZE>
Sets the per\-connection window size of HTTP/2 frontend
connection.
.sp
Default: \fB65535\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-http2\-window\-size=<SIZE>
Sets the initial window size of HTTP/2 backend
connection.
.sp
Default: \fB65535\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-http2\-connection\-window\-size=<SIZE>
Sets the per\-connection window size of HTTP/2 backend
connection.
.sp
Default: \fB2147483647\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-http2\-no\-cookie\-crumbling
Don\(aqt crumble cookie header field.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-padding=<N>
Add at most <N> bytes to a HTTP/2 frame payload as
padding. Specify 0 to disable padding. This option is
meant for debugging purpose and not intended to enhance
protocol security.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-no\-server\-push
Disable HTTP/2 server push. Server push is supported by
default mode and HTTP/2 frontend via Link header field.
It is also supported if both frontend and backend are
HTTP/2 in default mode. In this case, server push from
backend session is relayed to frontend, and server push
via Link header field is also supported.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http2\-optimize\-write\-buffer\-size
(Experimental) Enable write buffer size optimization in
frontend HTTP/2 TLS connection. This optimization aims
to reduce write buffer size so that it only contains
bytes which can send immediately. This makes server
more responsive to prioritized HTTP/2 stream because the
buffering of lower priority stream is reduced. This
option is only effective on recent Linux platform.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http2\-optimize\-window\-size
(Experimental) Automatically tune connection level
window size of frontend HTTP/2 TLS connection. If this
feature is enabled, connection window size starts with
the default window size, 65535 bytes. nghttpx
automatically adjusts connection window size based on
TCP receiving window size. The maximum window size is
capped by the value specified by
\fI\%\-\-frontend\-http2\-connection\-window\-size\fP\&. Since the
stream is subject to stream level window size, it should
be adjusted using \fI\%\-\-frontend\-http2\-window\-size\fP option as
well. This option is only effective on recent Linux
platform.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http2\-encoder\-dynamic\-table\-size=<SIZE>
Specify the maximum dynamic table size of HPACK encoder
in the frontend HTTP/2 connection. The decoder (client)
specifies the maximum dynamic table size it accepts.
Then the negotiated dynamic table size is the minimum of
this option value and the value which client specified.
.sp
Default: \fB4K\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http2\-decoder\-dynamic\-table\-size=<SIZE>
Specify the maximum dynamic table size of HPACK decoder
in the frontend HTTP/2 connection.
.sp
Default: \fB4K\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-http2\-encoder\-dynamic\-table\-size=<SIZE>
Specify the maximum dynamic table size of HPACK encoder
in the backend HTTP/2 connection. The decoder (backend)
specifies the maximum dynamic table size it accepts.
Then the negotiated dynamic table size is the minimum of
this option value and the value which backend specified.
.sp
Default: \fB4K\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-backend\-http2\-decoder\-dynamic\-table\-size=<SIZE>
Specify the maximum dynamic table size of HPACK decoder
in the backend HTTP/2 connection.
.sp
Default: \fB4K\fP
.UNINDENT
.SS Mode
.INDENT 0.0
.TP
.B (default mode)
Accept HTTP/2, and HTTP/1.1 over SSL/TLS. "no\-tls"
parameter is used in \fI\%\-\-frontend\fP option, accept HTTP/2
and HTTP/1.1 over cleartext TCP. The incoming HTTP/1.1
connection can be upgraded to HTTP/2 through HTTP
Upgrade.
.UNINDENT
.INDENT 0.0
.TP
.B \-s, \-\-http2\-proxy
Like default mode, but enable forward proxy. This is so
called HTTP/2 proxy mode.
.UNINDENT
.SS Logging
.INDENT 0.0
.TP
.B \-L, \-\-log\-level=<LEVEL>
Set the severity level of log output. <LEVEL> must be
one of INFO, NOTICE, WARN, ERROR and FATAL.
.sp
Default: \fBNOTICE\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-accesslog\-file=<PATH>
Set path to write access log. To reopen file, send USR1
signal to nghttpx.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-accesslog\-syslog
Send access log to syslog. If this option is used,
\fI\%\-\-accesslog\-file\fP option is ignored.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-accesslog\-format=<FORMAT>
Specify format string for access log. The default
format is combined format. The following variables are
available:
.INDENT 7.0
.IP \(bu 2
$remote_addr: client IP address.
.IP \(bu 2
$time_local: local time in Common Log format.
.IP \(bu 2
$time_iso8601: local time in ISO 8601 format.
.IP \(bu 2
$request: HTTP request line.
.IP \(bu 2
$status: HTTP response status code.
.IP \(bu 2
$body_bytes_sent: the number of bytes sent to client
as response body.
.IP \(bu 2
$http_<VAR>: value of HTTP request header <VAR> where
\(aq_\(aq in <VAR> is replaced with \(aq\-\(aq.
.IP \(bu 2
$remote_port: client port.
.IP \(bu 2
$server_port: server port.
.IP \(bu 2
$request_time: request processing time in seconds with
milliseconds resolution.
.IP \(bu 2
$pid: PID of the running process.
.IP \(bu 2
$alpn: ALPN identifier of the protocol which generates
the response. For HTTP/1, ALPN is always http/1.1,
regardless of minor version.
.IP \(bu 2
$tls_cipher: cipher used for SSL/TLS connection.
.IP \(bu 2
$tls_client_fingerprint_sha256: SHA\-256 fingerprint of
client certificate.
.IP \(bu 2
$tls_client_fingerprint_sha1: SHA\-1 fingerprint of
client certificate.
.IP \(bu 2
$tls_client_subject_name: subject name in client
certificate.
.IP \(bu 2
$tls_client_issuer_name: issuer name in client
certificate.
.IP \(bu 2
$tls_client_serial: serial number in client
certificate.
.IP \(bu 2
$tls_protocol: protocol for SSL/TLS connection.
.IP \(bu 2
$tls_session_id: session ID for SSL/TLS connection.
.IP \(bu 2
$tls_session_reused: "r" if SSL/TLS session was
reused. Otherwise, "."
.IP \(bu 2
$tls_sni: SNI server name for SSL/TLS connection.
.IP \(bu 2
$backend_host: backend host used to fulfill the
request. "\-" if backend host is not available.
.IP \(bu 2
$backend_port: backend port used to fulfill the
request. "\-" if backend host is not available.
.IP \(bu 2
$method: HTTP method
.IP \(bu 2
$path: Request path including query. For CONNECT
request, authority is recorded.
.IP \(bu 2
$path_without_query: $path up to the first \(aq?\(aq
character. For CONNECT request, authority is
recorded.
.IP \(bu 2
$protocol_version: HTTP version (e.g., HTTP/1.1,
HTTP/2)
.UNINDENT
.sp
The variable can be enclosed by "{" and "}" for
disambiguation (e.g., ${remote_addr}).
.sp
Default: \fB$remote_addr \- \- [$time_local] "$request" $status $body_bytes_sent "$http_referer" "$http_user_agent"\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-accesslog\-write\-early
Write access log when response header fields are
received from backend rather than when request
transaction finishes.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-errorlog\-file=<PATH>
Set path to write error log. To reopen file, send USR1
signal to nghttpx. stderr will be redirected to the
error log file unless \fI\%\-\-errorlog\-syslog\fP is used.
.sp
Default: \fB/dev/stderr\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-errorlog\-syslog
Send error log to syslog. If this option is used,
\fI\%\-\-errorlog\-file\fP option is ignored.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-syslog\-facility=<FACILITY>
Set syslog facility to <FACILITY>.
.sp
Default: \fBdaemon\fP
.UNINDENT
.SS HTTP
.INDENT 0.0
.TP
.B \-\-add\-x\-forwarded\-for
Append X\-Forwarded\-For header field to the downstream
request.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-strip\-incoming\-x\-forwarded\-for
Strip X\-Forwarded\-For header field from inbound client
requests.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-no\-add\-x\-forwarded\-proto
Don\(aqt append additional X\-Forwarded\-Proto header field
to the backend request. If inbound client sets
X\-Forwarded\-Proto, and
\fI\%\-\-no\-strip\-incoming\-x\-forwarded\-proto\fP option is used,
they are passed to the backend.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-no\-strip\-incoming\-x\-forwarded\-proto
Don\(aqt strip X\-Forwarded\-Proto header field from inbound
client requests.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-add\-forwarded=<LIST>
Append RFC 7239 Forwarded header field with parameters
specified in comma delimited list <LIST>. The supported
parameters are "by", "for", "host", and "proto". By
default, the value of "by" and "for" parameters are
obfuscated string. See \fI\%\-\-forwarded\-by\fP and
\fI\%\-\-forwarded\-for\fP options respectively. Note that nghttpx
does not translate non\-standard X\-Forwarded\-* header
fields into Forwarded header field, and vice versa.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-strip\-incoming\-forwarded
Strip Forwarded header field from inbound client
requests.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-forwarded\-by=(obfuscated|ip|<VALUE>)
Specify the parameter value sent out with "by" parameter
of Forwarded header field. If "obfuscated" is given,
the string is randomly generated at startup. If "ip" is
given, the interface address of the connection,
including port number, is sent with "by" parameter. In
case of UNIX domain socket, "localhost" is used instead
of address and port. User can also specify the static
obfuscated string. The limitation is that it must start
with "_", and only consists of character set
[A\-Za\-z0\-9._\-], as described in RFC 7239.
.sp
Default: \fBobfuscated\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-forwarded\-for=(obfuscated|ip)
Specify the parameter value sent out with "for"
parameter of Forwarded header field. If "obfuscated" is
given, the string is randomly generated for each client
connection. If "ip" is given, the remote client address
of the connection, without port number, is sent with
"for" parameter. In case of UNIX domain socket,
"localhost" is used instead of address.
.sp
Default: \fBobfuscated\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-no\-via
Don\(aqt append to Via header field. If Via header field
is received, it is left unaltered.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-no\-strip\-incoming\-early\-data
Don\(aqt strip Early\-Data header field from inbound client
requests.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-no\-location\-rewrite
Don\(aqt rewrite location header field in default mode.
When \fI\%\-\-http2\-proxy\fP is used, location header field will
not be altered regardless of this option.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-host\-rewrite
Rewrite host and :authority header fields in default
mode. When \fI\%\-\-http2\-proxy\fP is used, these headers will
not be altered regardless of this option.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-altsvc=<PROTOID,PORT[,HOST,[ORIGIN[,PARAMS]]]>
Specify protocol ID, port, host and origin of
alternative service. <HOST>, <ORIGIN> and <PARAMS> are
optional. Empty <HOST> and <ORIGIN> are allowed and
they are treated as nothing is specified. They are
advertised in alt\-svc header field only in HTTP/1.1
frontend. This option can be used multiple times to
specify multiple alternative services.
Example: \fI\%\-\-altsvc\fP="h2,443,,,ma=3600; persist=1"
.UNINDENT
.INDENT 0.0
.TP
.B \-\-http2\-altsvc=<PROTOID,PORT[,HOST,[ORIGIN[,PARAMS]]]>
Just like \fI\%\-\-altsvc\fP option, but this altsvc is only sent
in HTTP/2 frontend.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-add\-request\-header=<HEADER>
Specify additional header field to add to request header
set. This option just appends header field and won\(aqt
replace anything already set. This option can be used
several times to specify multiple header fields.
Example: \fI\%\-\-add\-request\-header\fP="foo: bar"
.UNINDENT
.INDENT 0.0
.TP
.B \-\-add\-response\-header=<HEADER>
Specify additional header field to add to response
header set. This option just appends header field and
won\(aqt replace anything already set. This option can be
used several times to specify multiple header fields.
Example: \fI\%\-\-add\-response\-header\fP="foo: bar"
.UNINDENT
.INDENT 0.0
.TP
.B \-\-request\-header\-field\-buffer=<SIZE>
Set maximum buffer size for incoming HTTP request header
field list. This is the sum of header name and value in
bytes. If trailer fields exist, they are counted
towards this number.
.sp
Default: \fB64K\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-max\-request\-header\-fields=<N>
Set maximum number of incoming HTTP request header
fields. If trailer fields exist, they are counted
towards this number.
.sp
Default: \fB100\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-response\-header\-field\-buffer=<SIZE>
Set maximum buffer size for incoming HTTP response
header field list. This is the sum of header name and
value in bytes. If trailer fields exist, they are
counted towards this number.
.sp
Default: \fB64K\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-max\-response\-header\-fields=<N>
Set maximum number of incoming HTTP response header
fields. If trailer fields exist, they are counted
towards this number.
.sp
Default: \fB500\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-error\-page=(<CODE>|*)=<PATH>
Set file path to custom error page served when nghttpx
originally generates HTTP error status code <CODE>.
<CODE> must be greater than or equal to 400, and at most
599. If "*" is used instead of <CODE>, it matches all
HTTP status code. If error status code comes from
backend server, the custom error pages are not used.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-server\-name=<NAME>
Change server response header field value to <NAME>.
.sp
Default: \fBnghttpx\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-no\-server\-rewrite
Don\(aqt rewrite server header field in default mode. When
\fI\%\-\-http2\-proxy\fP is used, these headers will not be altered
regardless of this option.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-redirect\-https\-port=<PORT>
Specify the port number which appears in Location header
field when redirect to HTTPS URI is made due to
"redirect\-if\-not\-tls" parameter in \fI\%\-\-backend\fP option.
.sp
Default: \fB443\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-require\-http\-scheme
Always require http or https scheme in HTTP request. It
also requires that https scheme must be used for an
encrypted connection. Otherwise, http scheme must be
used. This option is recommended for a server
deployment which directly faces clients and the services
it provides only require http or https scheme.
.UNINDENT
.SS API
.INDENT 0.0
.TP
.B \-\-api\-max\-request\-body=<SIZE>
Set the maximum size of request body for API request.
.sp
Default: \fB32M\fP
.UNINDENT
.SS DNS
.INDENT 0.0
.TP
.B \-\-dns\-cache\-timeout=<DURATION>
Set duration that cached DNS results remain valid. Note
that nghttpx caches the unsuccessful results as well.
.sp
Default: \fB10s\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-dns\-lookup\-timeout=<DURATION>
Set timeout that DNS server is given to respond to the
initial DNS query. For the 2nd and later queries,
server is given time based on this timeout, and it is
scaled linearly.
.sp
Default: \fB5s\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-dns\-max\-try=<N>
Set the number of DNS query before nghttpx gives up name
lookup.
.sp
Default: \fB2\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-max\-requests=<N>
The number of requests that single frontend connection
can process. For HTTP/2, this is the number of streams
in one HTTP/2 connection. For HTTP/1, this is the
number of keep alive requests. This is hint to nghttpx,
and it may allow additional few requests. The default
value is unlimited.
.UNINDENT
.SS Debug
.INDENT 0.0
.TP
.B \-\-frontend\-http2\-dump\-request\-header=<PATH>
Dumps request headers received by HTTP/2 frontend to the
file denoted in <PATH>. The output is done in HTTP/1
header field format and each header block is followed by
an empty line. This option is not thread safe and MUST
NOT be used with option \fI\%\-n\fP<N>, where <N> >= 2.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http2\-dump\-response\-header=<PATH>
Dumps response headers sent from HTTP/2 frontend to the
file denoted in <PATH>. The output is done in HTTP/1
header field format and each header block is followed by
an empty line. This option is not thread safe and MUST
NOT be used with option \fI\%\-n\fP<N>, where <N> >= 2.
.UNINDENT
.INDENT 0.0
.TP
.B \-o, \-\-frontend\-frame\-debug
Print HTTP/2 frames in frontend to stderr. This option
is not thread safe and MUST NOT be used with option
\fI\%\-n\fP=N, where N >= 2.
.UNINDENT
.SS Process
.INDENT 0.0
.TP
.B \-D, \-\-daemon
Run in a background. If \fI\%\-D\fP is used, the current working
directory is changed to \(aq\fI/\fP\(aq.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-pid\-file=<PATH>
Set path to save PID of this program.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-user=<USER>
Run this program as <USER>. This option is intended to
be used to drop root privileges.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-single\-process
Run this program in a single process mode for debugging
purpose. Without this option, nghttpx creates at least
2 processes: main and worker processes. If this option
is used, main and worker are unified into a single
process. nghttpx still spawns additional process if
neverbleed is used. In the single process mode, the
signal handling feature is disabled.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-max\-worker\-processes=<N>
The maximum number of worker processes. nghttpx spawns
new worker process when it reloads its configuration.
The previous worker process enters graceful termination
period and will terminate when it finishes handling the
existing connections. However, if reloading
configurations happen very frequently, the worker
processes might be piled up if they take a bit long time
to finish the existing connections. With this option,
if the number of worker processes exceeds the given
value, the oldest worker process is terminated
immediately. Specifying 0 means no limit and it is the
default behaviour.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-worker\-process\-grace\-shutdown\-period=<DURATION>
Maximum period for a worker process to terminate
gracefully. When a worker process enters in graceful
shutdown period (e.g., when nghttpx reloads its
configuration) and it does not finish handling the
existing connections in the given period of time, it is
immediately terminated. Specifying 0 means no limit and
it is the default behaviour.
.UNINDENT
.SS Scripting
.INDENT 0.0
.TP
.B \-\-mruby\-file=<PATH>
Set mruby script file
.UNINDENT
.INDENT 0.0
.TP
.B \-\-ignore\-per\-pattern\-mruby\-error
Ignore mruby compile error for per\-pattern mruby script
file. If error occurred, it is treated as if no mruby
file were specified for the pattern.
.UNINDENT
.SS HTTP/3 and QUIC
.INDENT 0.0
.TP
.B \-\-frontend\-quic\-idle\-timeout=<DURATION>
Specify an idle timeout for QUIC connection.
.sp
Default: \fB30s\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-quic\-debug\-log
Output QUIC debug log to \fI/dev/stderr.\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-quic\-bpf\-program\-file=<PATH>
Specify a path to eBPF program file reuseport_kern.o to
direct an incoming QUIC UDP datagram to a correct
socket.
.sp
Default: \fB/usr/local/lib/nghttp2/reuseport_kern.o\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-quic\-early\-data
Enable early data on frontend QUIC connections. nghttpx
sends "Early\-Data" header field to a backend server if a
request is received in early data and handshake has not
finished. All backend servers should deal with possibly
replayed requests.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-quic\-qlog\-dir=<DIR>
Specify a directory where a qlog file is written for
frontend QUIC connections. A qlog file is created per
each QUIC connection. The file name is ISO8601 basic
format, followed by "\-", server Source Connection ID and
".sqlog".
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-quic\-require\-token
Require an address validation token for a frontend QUIC
connection. Server sends a token in Retry packet or
NEW_TOKEN frame in the previous connection.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-quic\-congestion\-controller=<CC>
Specify a congestion controller algorithm for a frontend
QUIC connection. <CC> should be one of "cubic", "bbr",
and "bbr2".
.sp
Default: \fBcubic\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-quic\-secret\-file=<PATH>
Path to file that contains secure random data to be used
as QUIC keying materials. It is used to derive keys for
encrypting tokens and Connection IDs. It is not used to
encrypt QUIC packets. Each line of this file must
contain exactly 136 bytes hex\-encoded string (when
decoded the byte string is 68 bytes long). The first 2
bits of decoded byte string are used to identify the
keying material. An empty line or a line which starts
\(aq#\(aq is ignored. The file can contain more than one
keying materials. Because the identifier is 2 bits, at
most 4 keying materials are read and the remaining data
is discarded. The first keying material in the file is
primarily used for encryption and decryption for new
connection. The other ones are used to decrypt data for
the existing connections. Specifying multiple keying
materials enables key rotation. Please note that key
rotation does not occur automatically. User should
update files or change options values and restart
nghttpx gracefully. If opening or reading given file
fails, all loaded keying materials are discarded and it
is treated as if none of this option is given. If this
option is not given or an error occurred while opening
or reading a file, a keying material is generated
internally on startup and reload.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-quic\-server\-id=<HEXSTRING>
Specify server ID encoded in Connection ID to identify
this particular server instance. Connection ID is
encrypted and this part is not visible in public. It
must be 4 bytes long and must be encoded in hex string
(which is 8 bytes long). If this option is omitted, a
random server ID is generated on startup and
configuration reload.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-quic\-initial\-rtt=<DURATION>
Specify the initial RTT of the frontend QUIC connection.
.sp
Default: \fB333ms\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-no\-quic\-bpf
Disable eBPF.
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http3\-window\-size=<SIZE>
Sets the per\-stream initial window size of HTTP/3
frontend connection.
.sp
Default: \fB256K\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http3\-connection\-window\-size=<SIZE>
Sets the per\-connection window size of HTTP/3 frontend
connection.
.sp
Default: \fB1M\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http3\-max\-window\-size=<SIZE>
Sets the maximum per\-stream window size of HTTP/3
frontend connection. The window size is adjusted based
on the receiving rate of stream data. The initial value
is the value specified by \fI\%\-\-frontend\-http3\-window\-size\fP
and the window size grows up to <SIZE> bytes.
.sp
Default: \fB6M\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http3\-max\-connection\-window\-size=<SIZE>
Sets the maximum per\-connection window size of HTTP/3
frontend connection. The window size is adjusted based
on the receiving rate of stream data. The initial value
is the value specified by
\fI\%\-\-frontend\-http3\-connection\-window\-size\fP and the window
size grows up to <SIZE> bytes.
.sp
Default: \fB8M\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-frontend\-http3\-max\-concurrent\-streams=<N>
Set the maximum number of the concurrent streams in one
frontend HTTP/3 connection.
.sp
Default: \fB100\fP
.UNINDENT
.SS Misc
.INDENT 0.0
.TP
.B \-\-conf=<PATH>
Load configuration from <PATH>. Please note that
nghttpx always tries to read the default configuration
file if \fI\%\-\-conf\fP is not given.
.sp
Default: \fB/etc/nghttpx/nghttpx.conf\fP
.UNINDENT
.INDENT 0.0
.TP
.B \-\-include=<PATH>
Load additional configurations from <PATH>. File <PATH>
is read when configuration parser encountered this
option. This option can be used multiple times, or even
recursively.
.UNINDENT
.INDENT 0.0
.TP
.B \-v, \-\-version
Print version and exit.
.UNINDENT
.INDENT 0.0
.TP
.B \-h, \-\-help
Print this help and exit.
.UNINDENT
.sp
The <SIZE> argument is an integer and an optional unit (e.g., 10K is
10 * 1024). Units are K, M and G (powers of 1024).
.sp
The <DURATION> argument is an integer and an optional unit (e.g., 1s
is 1 second and 500ms is 500 milliseconds). Units are h, m, s or ms
(hours, minutes, seconds and milliseconds, respectively). If a unit
is omitted, a second is used as unit.
.SH FILES
.INDENT 0.0
.TP
.B \fI/etc/nghttpx/nghttpx.conf\fP
The default configuration file path nghttpx searches at startup.
The configuration file path can be changed using \fI\%\-\-conf\fP
option.
.sp
Those lines which are staring \fB#\fP are treated as comment.
.sp
The option name in the configuration file is the long command\-line
option name with leading \fB\-\-\fP stripped (e.g., \fBfrontend\fP). Put
\fB=\fP between option name and value. Don\(aqt put extra leading or
trailing spaces.
.sp
When specifying arguments including characters which have special
meaning to a shell, we usually use quotes so that shell does not
interpret them. When writing this configuration file, quotes for
this purpose must not be used. For example, specify additional
request header field, do this:
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
add\-request\-header=foo: bar
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
instead of:
.INDENT 7.0
.INDENT 3.5
.sp
.nf
.ft C
add\-request\-header="foo: bar"
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
The options which do not take argument in the command\-line \fItake\fP
argument in the configuration file. Specify \fByes\fP as an argument
(e.g., \fBhttp2\-proxy=yes\fP). If other string is given, it is
ignored.
.sp
To specify private key and certificate file which are given as
positional arguments in command\-line, use \fBprivate\-key\-file\fP and
\fBcertificate\-file\fP\&.
.sp
\fI\%\-\-conf\fP option cannot be used in the configuration file and
will be ignored if specified.
.TP
.B Error log
Error log is written to stderr by default. It can be configured
using \fI\%\-\-errorlog\-file\fP\&. The format of log message is as
follows:
.sp
<datetime> <main\-pid> <current\-pid> <thread\-id> <level> (<filename>:<line>) <msg>
.INDENT 7.0
.TP
.B <datetime>
It is a combination of date and time when the log is written. It
is in ISO 8601 format.
.TP
.B <main\-pid>
It is a main process ID.
.TP
.B <current\-pid>
It is a process ID which writes this log.
.TP
.B <thread\-id>
It is a thread ID which writes this log. It would be unique
within <current\-pid>.
.TP
.B <filename> and <line>
They are source file name, and line number which produce this log.
.TP
.B <msg>
It is a log message body.
.UNINDENT
.UNINDENT
.SH SIGNALS
.INDENT 0.0
.TP
.B SIGQUIT
Shutdown gracefully. First accept pending connections and stop
accepting connection. After all connections are handled, nghttpx
exits.
.TP
.B SIGHUP
Reload configuration file given in \fI\%\-\-conf\fP\&.
.TP
.B SIGUSR1
Reopen log files.
.UNINDENT
.sp
SIGUSR2
.INDENT 0.0
.INDENT 3.5
Fork and execute nghttpx. It will execute the binary in the same
path with same command\-line arguments and environment variables. As
of nghttpx version 1.20.0, the new main process sends SIGQUIT to the
original main process when it is ready to serve requests. For the
earlier versions of nghttpx, user has to send SIGQUIT to the
original main process.
.sp
The difference between SIGUSR2 (+ SIGQUIT) and SIGHUP is that former
is usually used to execute new binary, and the main process is newly
spawned. On the other hand, the latter just reloads configuration
file, and the same main process continues to exist.
.UNINDENT
.UNINDENT
.sp
\fBNOTE:\fP
.INDENT 0.0
.INDENT 3.5
nghttpx consists of multiple processes: one process for processing
these signals, and another one for processing requests. The former
spawns the latter. The former is called main process, and the
latter is called worker process. If neverbleed is enabled, the
worker process spawns neverbleed daemon process which does RSA key
processing. The above signal must be sent to the main process. If
the other processes received one of them, it is ignored. This
behaviour of these processes may change in the future release. In
other words, in the future release, the processes other than main
process may terminate upon the reception of these signals.
Therefore these signals should not be sent to the processes other
than main process.
.UNINDENT
.UNINDENT
.SH SERVER PUSH
.sp
nghttpx supports HTTP/2 server push in default mode with Link header
field. nghttpx looks for Link header field (\fI\%RFC 5988\fP) in response headers from
backend server and extracts URI\-reference with parameter
\fBrel=preload\fP (see \fI\%preload\fP)
and pushes those URIs to the frontend client. Here is a sample Link
header field to initiate server push:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
Link: </fonts/font.woff>; rel=preload
Link: </css/theme.css>; rel=preload
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Currently, the following restriction is applied for server push:
.INDENT 0.0
.IP 1. 3
The associated stream must have method "GET" or "POST". The
associated stream\(aqs status code must be 200.
.UNINDENT
.sp
This limitation may be loosened in the future release.
.sp
nghttpx also supports server push if both frontend and backend are
HTTP/2 in default mode. In this case, in addition to server push via
Link header field, server push from backend is forwarded to frontend
HTTP/2 session.
.sp
HTTP/2 server push will be disabled if \fI\%\-\-http2\-proxy\fP is
used.
.SH UNIX DOMAIN SOCKET
.sp
nghttpx supports UNIX domain socket with a filename for both frontend
and backend connections.
.sp
Please note that current nghttpx implementation does not delete a
socket with a filename. And on start up, if nghttpx detects that the
specified socket already exists in the file system, nghttpx first
deletes it. However, if SIGUSR2 is used to execute new binary and
both old and new configurations use same filename, new binary does not
delete the socket and continues to use it.
.SH OCSP STAPLING
.sp
OCSP query is done using external Python script
\fBfetch\-ocsp\-response\fP, which has been originally developed in Perl
as part of h2o project (\fI\%https://github.com/h2o/h2o\fP), and was
translated into Python.
.sp
The script file is usually installed under
\fB$(prefix)/share/nghttp2/\fP directory. The actual path to script can
be customized using \fI\%\-\-fetch\-ocsp\-response\-file\fP option.
.sp
If OCSP query is failed, previous OCSP response, if any, is continued
to be used.
.sp
\fI\%\-\-fetch\-ocsp\-response\-file\fP option provides wide range of
possibility to manage OCSP response. It can take an arbitrary script
or executable. The requirement is that it supports the command\-line
interface of \fBfetch\-ocsp\-response\fP script, and it must return a
valid DER encoded OCSP response on success. It must return exit code
0 on success, and 75 for temporary error, and the other error code for
generic failure. For large cluster of servers, it is not efficient
for each server to perform OCSP query using \fBfetch\-ocsp\-response\fP\&.
Instead, you can retrieve OCSP response in some way, and store it in a
disk or a shared database. Then specify a program in
\fI\%\-\-fetch\-ocsp\-response\-file\fP to fetch it from those stores.
This could provide a way to share the OCSP response between fleet of
servers, and also any OCSP query strategy can be applied which may be
beyond the ability of nghttpx itself or \fBfetch\-ocsp\-response\fP
script.
.SH TLS SESSION RESUMPTION
.sp
nghttpx supports TLS session resumption through both session ID and
session ticket.
.SS SESSION ID RESUMPTION
.sp
By default, session ID is shared by all worker threads.
.sp
If \fI\%\-\-tls\-session\-cache\-memcached\fP is given, nghttpx will
insert serialized session data to memcached with
\fBnghttpx:tls\-session\-cache:\fP + lowercase hex string of session ID
as a memcached entry key, with expiry time 12 hours. Session timeout
is set to 12 hours.
.sp
By default, connections to memcached server are not encrypted. To
enable encryption, use \fBtls\fP keyword in
\fI\%\-\-tls\-session\-cache\-memcached\fP option.
.SS TLS SESSION TICKET RESUMPTION
.sp
By default, session ticket is shared by all worker threads. The
automatic key rotation is also enabled by default. Every an hour, new
encryption key is generated, and previous encryption key becomes
decryption only key. We set session timeout to 12 hours, and thus we
keep at most 12 keys.
.sp
If \fI\%\-\-tls\-ticket\-key\-memcached\fP is given, encryption keys are
retrieved from memcached. nghttpx just reads keys from memcached; one
has to deploy key generator program to update keys frequently (e.g.,
every 1 hour). The example key generator tlsticketupdate.go is
available under contrib directory in nghttp2 archive. The memcached
entry key is \fBnghttpx:tls\-ticket\-key\fP\&. The data format stored in
memcached is the binary format described below:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
+\-\-\-\-\-\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-+
| VERSION (4) |LEN (2)|KEY(48 or 80) ...
+\-\-\-\-\-\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-+
^ |
| |
+\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-+
(LEN, KEY) pair can be repeated
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
All numbers in the above figure is bytes. All integer fields are
network byte order.
.sp
First 4 bytes integer VERSION field, which must be 1. The 2 bytes
integer LEN field gives the length of following KEY field, which
contains key. If \fI\%\-\-tls\-ticket\-key\-cipher\fP=aes\-128\-cbc is
used, LEN must be 48. If
\fI\%\-\-tls\-ticket\-key\-cipher\fP=aes\-256\-cbc is used, LEN must be
80. LEN and KEY pair can be repeated multiple times to store multiple
keys. The key appeared first is used as encryption key. All the
remaining keys are used as decryption only.
.sp
By default, connections to memcached server are not encrypted. To
enable encryption, use \fBtls\fP keyword in
\fI\%\-\-tls\-ticket\-key\-memcached\fP option.
.sp
If \fI\%\-\-tls\-ticket\-key\-file\fP is given, encryption key is read
from the given file. In this case, nghttpx does not rotate key
automatically. To rotate key, one has to restart nghttpx (see
SIGNALS).
.SH CERTIFICATE TRANSPARENCY
.sp
nghttpx supports TLS \fBsigned_certificate_timestamp\fP extension (\fI\%RFC
6962\fP). The relevant options
are \fI\%\-\-tls\-sct\-dir\fP and \fBsct\-dir\fP parameter in
\fI\%\-\-subcert\fP\&. They takes a directory, and nghttpx reads all
files whose extension is \fB\&.sct\fP under the directory. The \fB*.sct\fP
files are encoded as \fBSignedCertificateTimestamp\fP struct described
in \fI\%section 3.2 of RFC 69662\fP\&. This format is
the same one used by \fI\%nginx\-ct\fP and \fI\%mod_ssl_ct\fP\&.
\fI\%ct\-submit\fP can be
used to submit certificates to log servers, and obtain the
\fBSignedCertificateTimestamp\fP struct which can be used with nghttpx.
.SH MRUBY SCRIPTING
.sp
\fBWARNING:\fP
.INDENT 0.0
.INDENT 3.5
The current mruby extension API is experimental and not frozen. The
API is subject to change in the future release.
.UNINDENT
.UNINDENT
.sp
\fBWARNING:\fP
.INDENT 0.0
.INDENT 3.5
Almost all string value returned from method, or attribute is a
fresh new mruby string, which involves memory allocation, and
copies. Therefore, it is strongly recommended to store a return
value in a local variable, and use it, instead of calling method or
accessing attribute repeatedly.
.UNINDENT
.UNINDENT
.sp
nghttpx allows users to extend its capability using mruby scripts.
nghttpx has 2 hook points to execute mruby script: request phase and
response phase. The request phase hook is invoked after all request
header fields are received from client. The response phase hook is
invoked after all response header fields are received from backend
server. These hooks allows users to modify header fields, or common
HTTP variables, like authority or request path, and even return custom
response without forwarding request to backend servers.
.sp
There are 2 levels of mruby script invocations: global and
per\-pattern. The global mruby script is set by \fI\%\-\-mruby\-file\fP
option and is called for all requests. The per\-pattern mruby script
is set by "mruby" parameter in \fI\%\-b\fP option. It is invoked for
a request which matches the particular pattern. The order of hook
invocation is: global request phase hook, per\-pattern request phase
hook, per\-pattern response phase hook, and finally global response
phase hook. If a hook returns a response, any later hooks are not
invoked. The global request hook is invoked before the pattern
matching is made and changing request path may affect the pattern
matching.
.sp
Please note that request and response hooks of per\-pattern mruby
script for a single request might not come from the same script. This
might happen after a request hook is executed, backend failed for some
reason, and at the same time, backend configuration is replaced by API
request, and then the request uses new configuration on retry. The
response hook from new configuration, if it is specified, will be
invoked.
.sp
The all mruby script will be evaluated once per thread on startup, and
it must instantiate object and evaluate it as the return value (e.g.,
\fBApp.new\fP). This object is called app object. If app object
defines \fBon_req\fP method, it is called with \fI\%Nghttpx::Env\fP
object on request hook. Similarly, if app object defines \fBon_resp\fP
method, it is called with \fI\%Nghttpx::Env\fP object on response
hook. For each method invocation, user can can access
\fI\%Nghttpx::Request\fP and \fI\%Nghttpx::Response\fP objects
via \fI\%Nghttpx::Env#req\fP and \fI\%Nghttpx::Env#resp\fP
respectively.
.INDENT 0.0
.TP
.B Nghttpx::REQUEST_PHASE
Constant to represent request phase.
.UNINDENT
.INDENT 0.0
.TP
.B Nghttpx::RESPONSE_PHASE
Constant to represent response phase.
.UNINDENT
.INDENT 0.0
.TP
.B class Nghttpx::Env
Object to represent current request specific context.
.INDENT 7.0
.TP
.B attribute [R] req
Return \fI\%Request\fP object.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] resp
Return \fI\%Response\fP object.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] ctx
Return Ruby hash object. It persists until request finishes.
So values set in request phase hook can be retrieved in
response phase hook.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] phase
Return the current phase.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] remote_addr
Return IP address of a remote client. If connection is made
via UNIX domain socket, this returns the string "localhost".
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] server_addr
Return address of server that accepted the connection. This
is a string which specified in \fI\%\-\-frontend\fP option,
excluding port number, and not a resolved IP address. For
UNIX domain socket, this is a path to UNIX domain socket.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] server_port
Return port number of the server frontend which accepted the
connection from client.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_used
Return true if TLS is used on the connection.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_sni
Return the TLS SNI value which client sent in this connection.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_client_fingerprint_sha256
Return the SHA\-256 fingerprint of a client certificate.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_client_fingerprint_sha1
Return the SHA\-1 fingerprint of a client certificate.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_client_issuer_name
Return the issuer name of a client certificate.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_client_subject_name
Return the subject name of a client certificate.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_client_serial
Return the serial number of a client certificate.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_client_not_before
Return the start date of a client certificate in seconds since
the epoch.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_client_not_after
Return the end date of a client certificate in seconds since
the epoch.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_cipher
Return a TLS cipher negotiated in this connection.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_protocol
Return a TLS protocol version negotiated in this connection.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_session_id
Return a session ID for this connection in hex string.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_session_reused
Return true if, and only if a SSL/TLS session is reused.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] alpn
Return ALPN identifier negotiated in this connection.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] tls_handshake_finished
Return true if SSL/TLS handshake has finished. If it returns
false in the request phase hook, the request is received in
TLSv1.3 early data (0\-RTT) and might be vulnerable to the
replay attack. nghttpx will send Early\-Data header field to
backend servers to indicate this.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B class Nghttpx::Request
Object to represent request from client. The modification to
Request object is allowed only in request phase hook.
.INDENT 7.0
.TP
.B attribute [R] http_version_major
Return HTTP major version.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] http_version_minor
Return HTTP minor version.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R/W] method
HTTP method. On assignment, copy of given value is assigned.
We don\(aqt accept arbitrary method name. We will document them
later, but well known methods, like GET, PUT and POST, are all
supported.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R/W] authority
Authority (i.e., example.org), including optional port
component . On assignment, copy of given value is assigned.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R/W] scheme
Scheme (i.e., http, https). On assignment, copy of given
value is assigned.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R/W] path
Request path, including query component (i.e., /index.html).
On assignment, copy of given value is assigned. The path does
not include authority component of URI. This may include
query component. nghttpx makes certain normalization for
path. It decodes percent\-encoding for unreserved characters
(see \fI\%https://tools.ietf.org/html/rfc3986#section\-2.3\fP), and
resolves ".." and ".". But it may leave characters which
should be percent\-encoded as is. So be careful when comparing
path against desired string.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] headers
Return Ruby hash containing copy of request header fields.
Changing values in returned hash does not change request
header fields actually used in request processing. Use
\fI\%Nghttpx::Request#add_header\fP or
\fI\%Nghttpx::Request#set_header\fP to change request
header fields.
.UNINDENT
.INDENT 7.0
.TP
.B add_header(key, value)
Add header entry associated with key. The value can be single
string or array of string. It does not replace any existing
values associated with key.
.UNINDENT
.INDENT 7.0
.TP
.B set_header(key, value)
Set header entry associated with key. The value can be single
string or array of string. It replaces any existing values
associated with key.
.UNINDENT
.INDENT 7.0
.TP
.B clear_headers()
Clear all existing request header fields.
.UNINDENT
.INDENT 7.0
.TP
.B push(uri)
Initiate to push resource identified by \fIuri\fP\&. Only HTTP/2
protocol supports this feature. For the other protocols, this
method is noop. \fIuri\fP can be absolute URI, absolute path or
relative path to the current request. For absolute or
relative path, scheme and authority are inherited from the
current request. Currently, method is always GET. nghttpx
will issue request to backend servers to fulfill this request.
The request and response phase hooks will be called for pushed
resource as well.
.UNINDENT
.UNINDENT
.INDENT 0.0
.TP
.B class Nghttpx::Response
Object to represent response from backend server.
.INDENT 7.0
.TP
.B attribute [R] http_version_major
Return HTTP major version.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] http_version_minor
Return HTTP minor version.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R/W] status
HTTP status code. It must be in the range [200, 999],
inclusive. The non\-final status code is not supported in
mruby scripting at the moment.
.UNINDENT
.INDENT 7.0
.TP
.B attribute [R] headers
Return Ruby hash containing copy of response header fields.
Changing values in returned hash does not change response
header fields actually used in response processing. Use
\fI\%Nghttpx::Response#add_header\fP or
\fI\%Nghttpx::Response#set_header\fP to change response
header fields.
.UNINDENT
.INDENT 7.0
.TP
.B add_header(key, value)
Add header entry associated with key. The value can be single
string or array of string. It does not replace any existing
values associated with key.
.UNINDENT
.INDENT 7.0
.TP
.B set_header(key, value)
Set header entry associated with key. The value can be single
string or array of string. It replaces any existing values
associated with key.
.UNINDENT
.INDENT 7.0
.TP
.B clear_headers()
Clear all existing response header fields.
.UNINDENT
.INDENT 7.0
.TP
.B return(body)
Return custom response \fIbody\fP to a client. When this method
is called in request phase hook, the request is not forwarded
to the backend, and response phase hook for this request will
not be invoked. When this method is called in response phase
hook, response from backend server is canceled and discarded.
The status code and response header fields should be set
before using this method. To set status code, use
\fI\%Nghttpx::Response#status\fP\&. If status code is not
set, 200 is used. To set response header fields,
\fI\%Nghttpx::Response#add_header\fP and
\fI\%Nghttpx::Response#set_header\fP\&. When this method is
invoked in response phase hook, the response headers are
filled with the ones received from backend server. To send
completely custom header fields, first call
\fI\%Nghttpx::Response#clear_headers\fP to erase all
existing header fields, and then add required header fields.
It is an error to call this method twice for a given request.
.UNINDENT
.INDENT 7.0
.TP
.B send_info(status, headers)
Send non\-final (informational) response to a client. \fIstatus\fP
must be in the range [100, 199], inclusive. \fIheaders\fP is a
hash containing response header fields. Its key must be a
string, and the associated value must be either string or
array of strings. Since this is not a final response, even if
this method is invoked, request is still forwarded to a
backend unless \fI\%Nghttpx::Response#return\fP is called.
This method can be called multiple times. It cannot be called
after \fI\%Nghttpx::Response#return\fP is called.
.UNINDENT
.UNINDENT
.SS MRUBY EXAMPLES
.sp
Modify request path:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
class App
def on_req(env)
env.req.path = "/apps#{env.req.path}"
end
end
App.new
.ft P
.fi
.UNINDENT
.UNINDENT
.sp
Don\(aqt forget to instantiate and evaluate object at the last line.
.sp
Restrict permission of viewing a content to a specific client
addresses:
.INDENT 0.0
.INDENT 3.5
.sp
.nf
.ft C
class App
def on_req(env)
allowed_clients = ["127.0.0.1", "::1"]
if env.req.path.start_with?("/log/") &&
!allowed_clients.include?(env.remote_addr) then
env.resp.status = 404
env.resp.return "permission denied"
end
end
end
App.new
.ft P
.fi
.UNINDENT
.UNINDENT
.SH API ENDPOINTS
.sp
nghttpx exposes API endpoints to manipulate it via HTTP based API. By
default, API endpoint is disabled. To enable it, add a dedicated
frontend for API using \fI\%\-\-frontend\fP option with "api"
parameter. All requests which come from this frontend address, will
be treated as API request.
.sp
The response is normally JSON dictionary, and at least includes the
following keys:
.INDENT 0.0
.TP
.B status
The status of the request processing. The following values are
defined:
.INDENT 7.0
.TP
.B Success
The request was successful.
.TP
.B Failure
The request was failed. No change has been made.
.UNINDENT
.TP
.B code
HTTP status code
.UNINDENT
.sp
Additionally, depending on the API endpoint, \fBdata\fP key may be
present, and its value contains the API endpoint specific data.
.sp
We wrote "normally", since nghttpx may return ordinal HTML response in
some cases where the error has occurred before reaching API endpoint
(e.g., header field is too large).
.sp
The following section describes available API endpoints.
.SS POST /api/v1beta1/backendconfig
.sp
This API replaces the current backend server settings with the
requested ones. The request method should be POST, but PUT is also
acceptable. The request body must be nghttpx configuration file
format. For configuration file format, see \fI\%FILES\fP section. The
line separator inside the request body must be single LF (0x0A).
Currently, only \fI\%backend\fP option is parsed, the
others are simply ignored. The semantics of this API is replace the
current backend with the backend options in request body. Describe
the desired set of backend severs, and nghttpx makes it happen. If
there is no \fI\%backend\fP option is found in request
body, the current set of backend is replaced with the \fI\%backend\fP option\(aqs default value, which is \fB127.0.0.1,80\fP\&.
.sp
The replacement is done instantly without breaking existing
connections or requests. It also avoids any process creation as is
the case with hot swapping with signals.
.sp
The one limitation is that only numeric IP address is allowed in
\fI\%backend\fP in request body unless "dns" parameter
is used while non numeric hostname is allowed in command\-line or
configuration file is read using \fI\%\-\-conf\fP\&.
.SS GET /api/v1beta1/configrevision
.sp
This API returns configuration revision of the current nghttpx. The
configuration revision is opaque string, and it changes after each
reloading by SIGHUP. With this API, an external application knows
that whether nghttpx has finished reloading its configuration by
comparing the configuration revisions between before and after
reloading. It is recommended to disable persistent (keep\-alive)
connection for this purpose in order to avoid to send a request using
the reused connection which may bound to an old process.
.sp
This API returns response including \fBdata\fP key. Its value is JSON
object, and it contains at least the following key:
.INDENT 0.0
.TP
.B configRevision
The configuration revision of the current nghttpx
.UNINDENT
.SH SEE ALSO
.sp
\fBnghttp(1)\fP, \fBnghttpd(1)\fP, \fBh2load(1)\fP
.SH AUTHOR
Tatsuhiro Tsujikawa
.SH COPYRIGHT
2012, 2015, 2016, Tatsuhiro Tsujikawa
.\" Generated by docutils manpage writer.
.