ngx_http_lua_module - Embed the power of Lua into Nginx HTTP Servers.
This module is not distributed with the Nginx source. See the installation instructions.
Production ready.
This document describes ngx_lua v0.10.2 released on 8 March 2016.
# set search paths for pure Lua external libraries (';;' is the default path):
lua_package_path '/foo/bar/?.lua;/blah/?.lua;;';
# set search paths for Lua external libraries written in C (can also use ';;'):
lua_package_cpath '/bar/baz/?.so;/blah/blah/?.so;;';
server {
location /lua_content {
# MIME type determined by default_type:
default_type 'text/plain';
content_by_lua_block {
ngx.say('Hello,world!')
}
}
location /nginx_var {
# MIME type determined by default_type:
default_type 'text/plain';
# try access /nginx_var?a=hello,world
content_by_lua_block {
ngx.say(ngx.var.arg_a)
}
}
location = /request_body {
client_max_body_size 50k;
client_body_buffer_size 50k;
content_by_lua_block {
ngx.req.read_body() -- explicitly read the req body
local data = ngx.req.get_body_data()
if data then
ngx.say("body data:")
ngx.print(data)
return
end
-- body may get buffered in a temp file:
local file = ngx.req.get_body_file()
if file then
ngx.say("body is in file ", file)
else
ngx.say("no body found")
end
}
}
# transparent non-blocking I/O in Lua via subrequests
# (well, a better way is to use cosockets)
location = /lua {
# MIME type determined by default_type:
default_type 'text/plain';
content_by_lua_block {
local res = ngx.location.capture("/some_other_location")
if res then
ngx.say("status: ", res.status)
ngx.say("body:")
ngx.print(res.body)
end
}
}
location = /foo {
rewrite_by_lua_block {
res = ngx.location.capture("/memc",
{ args = { cmd = "incr", key = ngx.var.uri } }
)
}
proxy_pass http://blah.blah.com;
}
location = /mixed {
rewrite_by_lua_file /path/to/rewrite.lua;
access_by_lua_file /path/to/access.lua;
content_by_lua_file /path/to/content.lua;
}
# use nginx var in code path
# WARNING: contents in nginx var must be carefully filtered,
# otherwise there'll be great security risk!
location ~ ^/app/([-_a-zA-Z0-9/]+) {
set $path $1;
content_by_lua_file /path/to/lua/app/root/$path.lua;
}
location / {
client_max_body_size 100k;
client_body_buffer_size 100k;
access_by_lua_block {
-- check the client IP address is in our black list
if ngx.var.remote_addr == "132.5.72.3" then
ngx.exit(ngx.HTTP_FORBIDDEN)
end
-- check if the URI contains bad words
if ngx.var.uri and
string.match(ngx.var.request_body, "evil")
then
return ngx.redirect("/terms_of_use.html")
end
-- tests passed
}
# proxy_pass/fastcgi_pass/etc settings
}
}
Back to TOC
This module embeds Lua, via the standard Lua 5.1 interpreter or LuaJIT 2.0/2.1, into Nginx and by leveraging Nginx's subrequests, allows the integration of the powerful Lua threads (Lua coroutines) into the Nginx event model.
Unlike Apache's mod_lua and Lighttpd's mod_magnet, Lua code executed using this module can be 100% non-blocking on network traffic as long as the Nginx API for Lua provided by this module is used to handle requests to upstream services such as MySQL, PostgreSQL, Memcached, Redis, or upstream HTTP web services.
At least the following Lua libraries and Nginx modules can be used with this ngx_lua module:
Almost all the Nginx modules can be used with this ngx_lua module by means of ngx.location.capture orngx.location.capture_multi but it is recommended to use those lua-resty-*
libraries instead of creating subrequests to access the Nginx upstream modules because the former is usually much more flexible and memory-efficient.
The Lua interpreter or LuaJIT instance is shared across all the requests in a single nginx worker process but request contexts are segregated using lightweight Lua coroutines.
Loaded Lua modules persist in the nginx worker process level resulting in a small memory footprint in Lua even when under heavy loads.
This module is plugged into NGINX's "http" subsystem so it can only speaks downstream communication protocols in the HTTP family (HTTP 0.9/1.0/1.1/2.0, WebSockets, and etc). If you want to do generic TCP communications with the downstream clients, then you should use the ngx_stream_lua module instead which has a compatible Lua API.
Back to TOC
Just to name a few:
The possibilities are unlimited as the module allows bringing together various elements within Nginx as well as exposing the power of the Lua language to the user. The module provides the full flexibility of scripting while offering performance levels comparable with native C language programs both in terms of CPU time as well as memory footprint. This is particularly the case when LuaJIT 2.x is enabled.
Other scripting language implementations typically struggle to match this performance level.
The Lua state (Lua VM instance) is shared across all the requests handled by a single nginx worker process to minimize memory use.
Back to TOC
The latest version of this module is compatible with the following versions of Nginx:
Nginx cores older than 1.6.0 (exclusive) are not supported.
Back to TOC
It is highly recommended to use the OpenResty bundle that bundles Nginx, ngx_lua, LuaJIT 2.0/2.1 (or the optional standard Lua 5.1 interpreter), as well as a package of powerful companion Nginx modules. The basic installation step is a simple command:./configure --with-luajit && make && make install
.
Alternatively, ngx_lua can be manually compiled into Nginx:
Build the source with this module:
wget 'http://nginx.org/download/nginx-1.9.7.tar.gz'
tar -xzvf nginx-1.9.7.tar.gz
cd nginx-1.9.7/
# tell nginx's build system where to find LuaJIT 2.0:
export LUAJIT_LIB=/path/to/luajit/lib
export LUAJIT_INC=/path/to/luajit/include/luajit-2.0
# tell nginx's build system where to find LuaJIT 2.1:
export LUAJIT_LIB=/path/to/luajit/lib
export LUAJIT_INC=/path/to/luajit/include/luajit-2.1
# or tell where to find Lua if using Lua instead:
#export LUA_LIB=/path/to/lua/lib
#export LUA_INC=/path/to/lua/include
# Here we assume Nginx is to be installed under /opt/nginx/.
./configure --prefix=/opt/nginx \
--with-ld-opt="-Wl,-rpath,/path/to/luajit-or-lua/lib" \
--add-module=/path/to/ngx_devel_kit \
--add-module=/path/to/lua-nginx-module
make -j2
make install
Back to TOC
Starting from NGINX 1.9.11, you can also compile this module as a dynamic module, by using the --add-dynamic-module=PATH
option instead of --add-module=PATH
on the ./configure
command line above. And then you can explicitly load the module in your nginx.conf
via the load_module directive, for example,
load_module /path/to/modules/ndk_http_module.so; # assuming NDK is built as a dynamic module too
load_module /path/to/modules/ngx_http_lua_module.so;
Back to TOC
While building this module either via OpenResty or with the NGINX core, you can define the following C macros via the C compiler options:
NGX_LUA_USE_ASSERT
When defined, will enable assertions in the ngx_lua C code base. Recommended for debugging or testing builds. It can introduce some (small) runtime overhead when enabled. This macro was first introduced in the v0.9.10
release.NGX_LUA_ABORT_AT_PANIC
When the Lua/LuaJIT VM panics, ngx_lua will instruct the current nginx worker process to quit gracefully by default. By specifying this C macro, ngx_lua will abort the current nginx worker process (which usually result in a core dump file) immediately. This option is useful for debugging VM panics. This option was first introduced in the v0.9.8
release.NGX_LUA_NO_FFI_API
Excludes pure C API functions for FFI-based Lua API for NGINX (as required by lua-resty-core, for example). Enabling this macro can make the resulting binary code size smaller. To enable one or more of these macros, just pass extra C compiler options to the ./configure
script of either NGINX or OpenResty. For instance,
./configure --with-cc-opt="-DNGX_LUA_USE_ASSERT -DNGX_LUA_ABORT_AT_PANIC"
Back to TOC
Note that it is recommended to use LuaJIT 2.0 or LuaJIT 2.1 instead of the standard Lua 5.1 interpreter wherever possible.
If the standard Lua 5.1 interpreter is required however, run the following command to install it from the Ubuntu repository:
apt-get install -y lua5.1 liblua5.1-0 liblua5.1-0-dev
Everything should be installed correctly, except for one small tweak.
Library name liblua.so
has been changed in liblua5.1 package, it only comes with liblua5.1.so
, which needs to be symlinked to /usr/lib
so it could be found during the configuration process.
ln -s /usr/lib/x86_64-linux-gnu/liblua5.1.so /usr/lib/liblua.so
Back to TOC
Back to TOC
The openresty-en mailing list is for English speakers.
Back to TOC
The openresty mailing list is for Chinese speakers.
Back to TOC
The code repository of this project is hosted on github at openresty/lua-nginx-module.
Back to TOC
Please submit bug reports, wishlists, or patches by
Back to TOC
As from the v0.5.0rc32
release, all *_by_lua_file
configure directives (such as content_by_lua_file) support loading Lua 5.1 and LuaJIT 2.0/2.1 raw bytecode files directly.
Please note that the bytecode format used by LuaJIT 2.0/2.1 is not compatible with that used by the standard Lua 5.1 interpreter. So if using LuaJIT 2.0/2.1 with ngx_lua, LuaJIT compatible bytecode files must be generated as shown:
/path/to/luajit/bin/luajit -b /path/to/input_file.lua /path/to/output_file.luac
The -bg
option can be used to include debug information in the LuaJIT bytecode file:
/path/to/luajit/bin/luajit -bg /path/to/input_file.lua /path/to/output_file.luac
Please refer to the official LuaJIT documentation on the -b
option for more details:
http://luajit.org/running.html#opt_b
Also, the bytecode files generated by LuaJIT 2.1 is not compatible with LuaJIT 2.0, and vice versa. The support for LuaJIT 2.1 bytecode was first added in ngx_lua v0.9.3.
Similarly, if using the standard Lua 5.1 interpreter with ngx_lua, Lua compatible bytecode files must be generated using the luac
commandline utility as shown:
luac -o /path/to/output_file.luac /path/to/input_file.lua
Unlike as with LuaJIT, debug information is included in standard Lua 5.1 bytecode files by default. This can be striped out by specifying the -s
option as shown:
luac -s -o /path/to/output_file.luac /path/to/input_file.lua
Attempts to load standard Lua 5.1 bytecode files into ngx_lua instances linked to LuaJIT 2.0/2.1 or vice versa, will result in an error message, such as that below, being logged into the Nginx error.log
file:
[error] 13909#0: *1 failed to load Lua inlined code: bad byte-code header in /path/to/test_file.luac
Loading bytecode files via the Lua primitives like require
and dofile
should always work as expected.
Back to TOC
If you want to access the system environment variable, say, foo
, in Lua via the standard Lua API os.getenv, then you should also list this environment variable name in your nginx.conf
file via the env directive. For example,
env foo;
Back to TOC
The HTTP 1.0 protocol does not support chunked output and requires an explicit Content-Length
header when the response body is not empty in order to support the HTTP 1.0 keep-alive. So when a HTTP 1.0 request is made and thelua_http10_buffering directive is turned on
, ngx_lua will buffer the output of ngx.say and ngx.print calls and also postpone sending response headers until all the response body output is received. At that time ngx_lua can calculate the total length of the body and construct a proper Content-Length
header to return to the HTTP 1.0 client. If the Content-Length
response header is set in the running Lua code, however, this buffering will be disabled even if the lua_http10_buffering directive is turned on
.
For large streaming output responses, it is important to disable the lua_http10_buffering directive to minimise memory usage.
Note that common HTTP benchmark tools such as ab
and http_load
issue HTTP 1.0 requests by default. To force curl
to send HTTP 1.0 requests, use the -0
option.
Back to TOC
When LuaJIT 2.x is used, it is possible to statically link the bytecode of pure Lua modules into the Nginx executable.
Basically you use the luajit
executable to compile .lua
Lua module files to .o
object files containing the exported bytecode data, and then link the .o
files directly in your Nginx build.
Below is a trivial example to demonstrate this. Consider that we have the following .lua
file named foo.lua
:
-- foo.lua
local _M = {}
function _M.go()
print("Hello from foo")
end
return _M
And then we compile this .lua
file to foo.o
file:
/path/to/luajit/bin/luajit -bg foo.lua foo.o
What matters here is the name of the .lua
file, which determines how you use this module later on the Lua land. The file namefoo.o
does not matter at all except the .o
file extension (which tells luajit
what output format is used). If you want to strip the Lua debug information from the resulting bytecode, you can just specify the -b
option above instead of -bg
.
Then when building Nginx or OpenResty, pass the --with-ld-opt="foo.o"
option to the ./configure
script:
./configure --with-ld-opt="/path/to/foo.o" ...
Finally, you can just do the following in any Lua code run by ngx_lua:
local foo = require "foo"
foo.go()
And this piece of code no longer depends on the external foo.lua
file any more because it has already been compiled into thenginx
executable.
If you want to use dot in the Lua module name when calling require
, as in
local foo = require "resty.foo"
then you need to rename the foo.lua
file to resty_foo.lua
before compiling it down to a .o
file with the luajit
command-line utility.
It is important to use exactly the same version of LuaJIT when compiling .lua
files to .o
files as building nginx + ngx_lua. This is because the LuaJIT bytecode format may be incompatible between different LuaJIT versions. When the bytecode format is incompatible, you will see a Lua runtime error saying that the Lua module is not found.
When you have multiple .lua
files to compile and link, then just specify their .o
files at the same time in the value of the --with-ld-opt
option. For instance,
./configure --with-ld-opt="/path/to/foo.o /path/to/bar.o" ...
If you have just too many .o
files, then it might not be feasible to name them all in a single command. In this case, you can build a static library (or archive) for your .o
files, as in
ar rcus libmyluafiles.a *.o
then you can link the myluafiles
archive as a whole to your nginx executable:
./configure \
--with-ld-opt="-L/path/to/lib -Wl,--whole-archive -lmyluafiles -Wl,--no-whole-archive"
where /path/to/lib
is the path of the directory containing the libmyluafiles.a
file. It should be noted that the linker option --whole-archive
is required here because otherwise our archive will be skipped because no symbols in our archive are mentioned in the main parts of the nginx executable.
Back to TOC
To globally share data among all the requests handled by the same nginx worker process, encapsulate the shared data into a Lua module, use the Lua require
builtin to import the module, and then manipulate the shared data in Lua. This works because required Lua modules are loaded only once and all coroutines will share the same copy of the module (both its code and data). Note however that Lua global variables (note, not module-level variables) WILL NOT persist between requests because of the one-coroutine-per-request isolation design.
Here is a complete small example:
-- mydata.lua
local _M = {}
local data = {
dog = 3,
cat = 4,
pig = 5,
}
function _M.get_age(name)
return data[name]
end
return _M
and then accessing it from nginx.conf
:
location /lua {
content_by_lua '
local mydata = require "mydata"
ngx.say(mydata.get_age("dog"))
';
}
The mydata
module in this example will only be loaded and run on the first request to the location /lua
, and all subsequent requests to the same nginx worker process will use the reloaded instance of the module as well as the same copy of the data in it, until a HUP
signal is sent to the Nginx master process to force a reload. This data sharing technique is essential for high performance Lua applications based on this module.
Note that this data sharing is on a per-worker basis and not on a per-server basis. That is, when there are multiple nginx worker processes under an Nginx master, data sharing cannot cross the process boundary between these workers.
It is usually recommended to share read-only data this way. You can also share changeable data among all the concurrent requests of each nginx worker process as long as there is no nonblocking I/O operations (including ngx.sleep) in the middle of your calculations. As long as you do not give the control back to the nginx event loop and ngx_lua's light thread scheduler (even implicitly), there can never be any race conditions in between. For this reason, always be very careful when you want to share changeable data on the worker level. Buggy optimizations can easily lead to hard-to-debug race conditions under load.
If server-wide data sharing is required, then use one or more of the following approaches:
memcached
, redis
, MySQL
or PostgreSQL
. The OpenResty bundle associated with this module comes with a set of companion Nginx modules and Lua libraries that provide interfaces with these data storage mechanisms.Back to TOC
Back to TOC
The tcpsock:connect method may indicate success
despite connection failures such as with Connection Refused
errors.
However, later attempts to manipulate the cosocket object will fail and return the actual error status message generated by the failed connect operation.
This issue is due to limitations in the Nginx event model and only appears to affect Mac OS X.
Back to TOC
dofile
and require
builtins are currently implemented as C functions in both Lua 5.1 and LuaJIT 2.0/2.1, if the Lua file being loaded by dofile
or require
invokes ngx.location.capture*, ngx.exec, ngx.exit, or other API functions requiring yielding in the top-level scope of the Lua file, then the Lua error "attempt to yield across C-call boundary" will be raised. To avoid this, put these calls requiring yielding into your own Lua functions in the Lua file instead of the top-level scope of the file.for ... in ...
statement when the standard Lua 5.1 interpreter is used and the attempt to yield across metamethod/C-call boundary
error will be produced. Please use LuaJIT 2.x, which supports a fully resumable VM, to avoid this.Back to TOC
Care must be taken when importing modules and this form should be used:
local xxx = require('xxx')
instead of the old deprecated form:
require('xxx')
Here is the reason: by design, the global environment has exactly the same lifetime as the Nginx request handler associated with it. Each request handler has its own set of Lua global variables and that is the idea of request isolation. The Lua module is actually loaded by the first Nginx request handler and is cached by the require()
built-in in the package.loaded
table for later reference, and the module()
builtin used by some Lua modules has the side effect of setting a global variable to the loaded module table. But this global variable will be cleared at the end of the request handler, and every subsequent request handler all has its own (clean) global environment. So one will get Lua exception for accessing the nil
value.
Generally, use of Lua global variables is a really really bad idea in the context of ngx_lua because
It's highly recommended to always declare them via "local" in the scope that is reasonable.
To find out all the uses of Lua global variables in your Lua code, you can run the lua-releng tool across all your .lua source files:
$ lua-releng
Checking use of Lua global variables in file lib/foo/bar.lua ...
1 [1489] SETGLOBAL 7 -1 ; contains
55 [1506] GETGLOBAL 7 -3 ; setvar
3 [1545] GETGLOBAL 3 -4 ; varexpand
The output says that the line 1489 of file lib/foo/bar.lua
writes to a global variable named contains
, the line 1506 reads from the global variable setvar
, and line 1545 reads the global varexpand
.
This tool will guarantee that local variables in the Lua module functions are all declared with the local
keyword, otherwise a runtime exception will be thrown. It prevents undesirable race conditions while accessing such variables. See Data Sharing within an Nginx Worker for the reasons behind this.
Back to TOC
The ngx.location.capture and ngx.location.capture_multi directives cannot capture locations that include the add_before_body,add_after_body, auth_request, echo_location, echo_location_async, echo_subrequest, or echo_subrequest_async directives.
location /foo {
content_by_lua '
res = ngx.location.capture("/bar")
';
}
location /bar {
echo_location /blah;
}
location /blah {
echo "Success!";
}
$ curl -i http://example.com/foo
will not work as expected.
Back to TOC
Due the internal limitations in the nginx core, the cosocket API are disabled in the following contexts: set_by_lua*, log_by_lua*,header_filter_by_lua*, and body_filter_by_lua.
The cosockets are currently also disabled in the init_by_lua* and init_worker_by_lua* directive contexts but we may add support for these contexts in the future because there is no limitation in the nginx core (or the limitation might be worked around).
There exists a work-around, however, when the original context does not need to wait for the cosocket results. That is, creating a 0-delay timer via the ngx.timer.at API and do the cosocket results in the timer handler, which runs asynchronously as to the original context creating the timer.
Back to TOC
WARNING We no longer suffer from this pitfall since the introduction of the *_by_lua_block {}
configuration directives.
PCRE sequences such as \d
, \s
, or \w
, require special attention because in string literals, the backslash character, \
, is stripped out by both the Lua language parser and by the Nginx config file parser before processing. So the following snippet will not work as expected:
# nginx.conf
? location /test {
? content_by_lua '
? local regex = "\d+" -- THIS IS WRONG!!
? local m = ngx.re.match("hello, 1234", regex)
? if m then ngx.say(m[0]) else ngx.say("not matched!") end
? ';
? }
# evaluates to "not matched!"
To avoid this, double escape the backslash:
# nginx.conf
location /test {
content_by_lua '
local regex = "\\\\d+"
local m = ngx.re.match("hello, 1234", regex)
if m then ngx.say(m[0]) else ngx.say("not matched!") end
';
}
# evaluates to "1234"
Here, \\\\d+
is stripped down to \\d+
by the Nginx config file parser and this is further stripped down to \d+
by the Lua language parser before running.
Alternatively, the regex pattern can be presented as a long-bracketed Lua string literal by encasing it in "long brackets",[[...]]
, in which case backslashes have to only be escaped once for the Nginx config file parser.
# nginx.conf
location /test {
content_by_lua '
local regex = [[\\d+]]
local m = ngx.re.match("hello, 1234", regex)
if m then ngx.say(m[0]) else ngx.say("not matched!") end
';
}
# evaluates to "1234"
Here, [[\\d+]]
is stripped down to [[\d+]]
by the Nginx config file parser and this is processed correctly.
Note that a longer from of the long bracket, [=[...]=]
, may be required if the regex pattern contains [...]
sequences. The[=[...]=]
form may be used as the default form if desired.
# nginx.conf
location /test {
content_by_lua '
local regex = [=[[0-9]+]=]
local m = ngx.re.match("hello, 1234", regex)
if m then ngx.say(m[0]) else ngx.say("not matched!") end
';
}
# evaluates to "1234"
An alternative approach to escaping PCRE sequences is to ensure that Lua code is placed in external script files and executed using the various *_by_lua_file
directives. With this approach, the backslashes are only stripped by the Lua language parser and therefore only need to be escaped once each.
-- test.lua
local regex = "\\d+"
local m = ngx.re.match("hello, 1234", regex)
if m then ngx.say(m[0]) else ngx.say("not matched!") end
-- evaluates to "1234"
Within external script files, PCRE sequences presented as long-bracketed Lua string literals do not require modification.
-- test.lua
local regex = [[\d+]]
local m = ngx.re.match("hello, 1234", regex)
if m then ngx.say(m[0]) else ngx.say("not matched!") end
-- evaluates to "1234"
Back to TOC
Mixing SSI with ngx_lua in the same Nginx request is not supported at all. Just use ngx_lua exclusively. Everything you can do with SSI can be done atop ngx_lua anyway and it can be more efficient when using ngx_lua.
Back to TOC
Certain Lua APIs provided by ngx_lua do not work in Nginx's SPDY mode yet: ngx.location.capture, ngx.location.capture_multi, and ngx.req.socket.
Back to TOC
Nginx may terminate a request early with (at least):
This means that phases that normally run are skipped, such as the rewrite or access phase. This also means that later phases that are run regardless, e.g. log_by_lua, will not have access to information that is normally set in those phases.
Back to TOC
datagram {
server {
listen 1953;
handler_by_lua '
-- custom Lua code implementing the special UDP server...
';
}
}
bind()
method for stream-typed cosockets.connect
queueing when the backend concurrency exceeds its connection pool limit.bsdrecv
method.ngx.resp.add_header
to emulate the standard add_header
config directive.ngx.re.split()
.extra_headers
optionngx_hash_t
to optimize the built-in header look-up process for ngx.req.set_header, ngx.header.HEADER, and etc.ignore_resp_headers
, ignore_resp_body
, and ignore_resp
options to ngx.location.capture andngx.location.capture_multi methods, to allow micro performance tuning on the user side.stat
mode similar to mod_lua.Back to TOC
The changes of every release of this module can be obtained from the OpenResty bundle's change logs:
http://openresty.org/#Changes
Back to TOC
The following dependencies are required to run the test suite:
Nginx version >= 1.4.2
Perl modules:
Nginx modules:
The order in which these modules are added during configuration is important because the position of any filter module in the filtering chain determines the final output, for example. The correct adding order is shown above.
3rd-party Lua libraries:
Applications:
See also the developer build script for more details on setting up the testing environment.
To run the whole test suite in the default testing mode:
cd /path/to/lua-nginx-module
export PATH=/path/to/your/nginx/sbin:$PATH
prove -I/path/to/test-nginx/lib -r t
To run specific test files:
cd /path/to/lua-nginx-module
export PATH=/path/to/your/nginx/sbin:$PATH
prove -I/path/to/test-nginx/lib t/002-content.t t/003-errors.t
To run a specific test block in a particular test file, add the line --- ONLY
to the test block you want to run, and then use theprove
utility to run that .t
file.
There are also various testing modes based on mockeagain, valgrind, and etc. Refer to the Test::Nginx documentation for more details for various advanced testing modes. See also the test reports for the Nginx test cluster running on Amazon EC2:http://qa.openresty.org.
Back to TOC
This module is licensed under the BSD license.
Copyright (C) 2009-2016, by Xiaozhe Wang (chaoslawful) [email protected].
Copyright (C) 2009-2016, by Yichun "agentzh" Zhang (章亦春) [email protected], CloudFlare Inc.
All rights reserved.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
Back to TOC
Back to TOC
Back to TOC
syntax: lua_use_default_type on | off
default: lua_use_default_type on
context: http, server, location, location if
Specifies whether to use the MIME type specified by the default_type directive for the default value of the Content-Type
response header. If you do not want a default Content-Type
response header for your Lua request handlers, then turn this directive off.
This directive is turned on by default.
This directive was first introduced in the v0.9.1
release.
Back to TOC
syntax: lua_code_cache on | off
default: lua_code_cache on
context: http, server, location, location if
Enables or disables the Lua code cache for Lua code in *_by_lua_file
directives (like set_by_lua_file andcontent_by_lua_file) and Lua modules.
When turning off, every request served by ngx_lua will run in a separate Lua VM instance, starting from the 0.9.3
release. So the Lua files referenced in set_by_lua_file, content_by_lua_file, access_by_lua_file, and etc will not be cached and all Lua modules used will be loaded from scratch. With this in place, developers can adopt an edit-and-refresh approach.
Please note however, that Lua code written inlined within nginx.conf such as those specified by set_by_lua, content_by_lua,access_by_lua, and rewrite_by_lua will not be updated when you edit the inlined Lua code in your nginx.conf
file because only the Nginx config file parser can correctly parse the nginx.conf
file and the only way is to reload the config file by sending a HUP
signal or just to restart Nginx.
Even when the code cache is enabled, Lua files which are loaded by dofile
or loadfile
in *_by_lua_file cannot be cached (unless you cache the results yourself). Usually you can either use the init_by_lua or init_by_lua_file directives to load all such files or just make these Lua files true Lua modules and load them via require
.
The ngx_lua module does not support the stat
mode available with the Apache mod_lua
module (yet).
Disabling the Lua code cache is strongly discouraged for production use and should only be used during development as it has a significant negative impact on overall performance. For example, the performance a "hello world" Lua example can drop by an order of magnitude after disabling the Lua code cache.
Back to TOC
syntax: lua_regex_cache_max_entries
default: lua_regex_cache_max_entries 1024
context: http
Specifies the maximum number of entries allowed in the worker process level compiled regex cache.
The regular expressions used in ngx.re.match, ngx.re.gmatch, ngx.re.sub, and ngx.re.gsub will be cached within this cache if the regex option o
(i.e., compile-once flag) is specified.
The default number of entries allowed is 1024 and when this limit is reached, new regular expressions will not be cached (as if theo
option was not specified) and there will be one, and only one, warning in the error.log
file:
2011/08/27 23:18:26 [warn] 31997#0: *1 lua exceeding regex cache max entries (1024), ...
Do not activate the o
option for regular expressions (and/or replace
string arguments for ngx.re.sub and ngx.re.gsub) that are generated on the fly and give rise to infinite variations to avoid hitting the specified limit.
Back to TOC
syntax: lua_regex_match_limit
default: lua_regex_match_limit 0
context: http
Specifies the "match limit" used by the PCRE library when executing the ngx.re API. To quote the PCRE manpage, "the limit ... has the effect of limiting the amount of backtracking that can take place."
When the limit is hit, the error string "pcre_exec() failed: -8" will be returned by the ngx.re API functions on the Lua land.
When setting the limit to 0, the default "match limit" when compiling the PCRE library is used. And this is the default value of this directive.
This directive was first introduced in the v0.8.5
release.
Back to TOC
syntax: lua_package_path
default: The content of LUA_PATH environ variable or Lua's compiled-in defaults.
context: http
Sets the Lua module search path used by scripts specified by set_by_lua, content_by_lua and others. The path string is in standard Lua path form, and ;;
can be used to stand for the original search paths.
As from the v0.5.0rc29
release, the special notation $prefix
or ${prefix}
can be used in the search path string to indicate the path of the server prefix
usually determined by the -p PATH
command-line option while starting the Nginx server.
Back to TOC
syntax: lua_package_cpath
default: The content of LUA_CPATH environment variable or Lua's compiled-in defaults.
context: http
Sets the Lua C-module search path used by scripts specified by set_by_lua, content_by_lua and others. The cpath string is in standard Lua cpath form, and ;;
can be used to stand for the original cpath.
As from the v0.5.0rc29
release, the special notation $prefix
or ${prefix}
can be used in the search path string to indicate the path of the server prefix
usually determined by the -p PATH
command-line option while starting the Nginx server.
Back to TOC
syntax: init_by_lua
context: http
phase: loading-config
WARNING Since the v0.9.17
release, use of this directive is discouraged; use the new init_by_lua_block directive instead.
Runs the Lua code specified by the argument
on the global Lua VM level when the Nginx master process (if any) is loading the Nginx config file.
When Nginx receives the HUP
signal and starts reloading the config file, the Lua VM will also be re-created and init_by_lua
will run again on the new Lua VM. In case that the lua_code_cache directive is turned off (default on), the init_by_lua
handler will run upon every request because in this special mode a standalone Lua VM is always created for each request.
Usually you can register (true) Lua global variables or pre-load Lua modules at server start-up by means of this hook. Here is an example for pre-loading Lua modules:
init_by_lua 'cjson = require "cjson"';
server {
location = /api {
content_by_lua '
ngx.say(cjson.encode({dog = 5, cat = 6}))
';
}
}
You can also initialize the lua_shared_dict shm storage at this phase. Here is an example for this:
lua_shared_dict dogs 1m;
init_by_lua '
local dogs = ngx.shared.dogs;
dogs:set("Tom", 56)
';
server {
location = /api {
content_by_lua '
local dogs = ngx.shared.dogs;
ngx.say(dogs:get("Tom"))
';
}
}
But note that, the lua_shared_dict's shm storage will not be cleared through a config reload (via the HUP
signal, for example). So if you do not want to re-initialize the shm storage in your init_by_lua
code in this case, then you just need to set a custom flag in the shm storage and always check the flag in your init_by_lua
code.
Because the Lua code in this context runs before Nginx forks its worker processes (if any), data or code loaded here will enjoy theCopy-on-write (COW) feature provided by many operating systems among all the worker processes, thus saving a lot of memory.
Do not initialize your own Lua global variables in this context because use of Lua global variables have performance penalties and can lead to global namespace pollution (see the Lua Variable Scope section for more details). The recommended way is to use proper Lua module files (but do not use the standard Lua function module() to define Lua modules because it pollutes the global namespace as well) and call require() to load your own module files in init_by_lua
or other contexts (require() does cache the loaded Lua modules in the global package.loaded
table in the Lua registry so your modules will only loaded once for the whole Lua VM instance).
Only a small set of the Nginx API for Lua is supported in this context:
More Nginx APIs for Lua may be supported in this context upon future user requests.
Basically you can safely use Lua libraries that do blocking I/O in this very context because blocking the master process during server start-up is completely okay. Even the Nginx core does blocking I/O (at least on resolving upstream's host names) at the configure-loading phase.
You should be very careful about potential security vulnerabilities in your Lua code registered in this context because the Nginx master process is often run under the root
account.
This directive was first introduced in the v0.5.5
release.
Back to TOC
syntax: init_by_lua_block { lua-script }
context: http
phase: loading-config
Similar to the init_by_lua directive except that this directive inlines the Lua source directly inside a pair of curly braces ({}
) instead of in an NGINX string literal (which requires special character escaping).
For instance,
init_by_lua_block {
print("I need no extra escaping here, for example: \r\nblah")
}
This directive was first introduced in the v0.9.17
release.
Back to TOC
syntax: init_by_lua_file
context: http
phase: loading-config
Equivalent to init_by_lua, except that the file specified by
contains the Lua code or Lua/LuaJIT bytecode to be executed.
When a relative path like foo/bar.lua
is given, they will be turned into the absolute path relative to the server prefix
path determined by the -p PATH
command-line option while starting the Nginx server.
This directive was first introduced in the v0.5.5
release.
Back to TOC
syntax: init_worker_by_lua
context: http
phase: starting-worker
WARNING Since the v0.9.17
release, use of this directive is discouraged; use the new init_worker_by_lua_block directive instead.
Runs the specified Lua code upon every Nginx worker process's startup when the master process is enabled. When the master process is disabled, this hook will just run after init_by_lua*.
This hook is often used to create per-worker reoccurring timers (via the ngx.timer.at Lua API), either for backend health-check or other timed routine work. Below is an example,
init_worker_by_lua '
local delay = 3 -- in seconds
local new_timer = ngx.timer.at
local log = ngx.log
local ERR = ngx.ERR
local check
check = function(premature)
if not premature then
-- do the health check or other routine work
local ok, err = new_timer(delay, check)
if not ok then
log(ERR, "failed to create timer: ", err)
return
end
end
end
local ok, err = new_timer(delay, check)
if not ok then
log(ERR, "failed to create timer: ", err)
return
end
';
This directive was first introduced in the v0.9.5
release.
Back to TOC
syntax: init_worker_by_lua_block { lua-script }
context: http
phase: starting-worker
Similar to the init_worker_by_lua directive except that this directive inlines the Lua source directly inside a pair of curly braces ({}
) instead of in an NGINX string literal (which requires special character escaping).
For instance,
init_worker_by_lua_block {
print("I need no extra escaping here, for example: \r\nblah")
}
This directive was first introduced in the v0.9.17
release.
Back to TOC
syntax: init_worker_by_lua_file
context: http
phase: starting-worker
Similar to init_worker_by_lua, but accepts the file path to a Lua source file or Lua bytecode file.
This directive was first introduced in the v0.9.5
release.
Back to TOC
syntax: set_by_lua $res
context: server, server if, location, location if
phase: rewrite
WARNING Since the v0.9.17
release, use of this directive is discouraged; use the new set_by_lua_block directive instead.
Executes code specified in
with optional input arguments $arg1 $arg2 ...
, and returns string output to$res
. The code in
can make API calls and can retrieve input arguments from the ngx.arg
table (index starts from 1
and increases sequentially).
This directive is designed to execute short, fast running code blocks as the Nginx event loop is blocked during code execution. Time consuming code sequences should therefore be avoided.
This directive is implemented by injecting custom commands into the standard ngx_http_rewrite_module's command list. Because ngx_http_rewrite_module does not support nonblocking I/O in its commands, Lua APIs requiring yielding the current Lua "light thread" cannot work in this directive.
At least the following API functions are currently disabled within the context of set_by_lua
:
In addition, note that this directive can only write out a value to a single Nginx variable at a time. However, a workaround is possible using the ngx.var.VARIABLE interface.
location /foo {
set $diff ''; # we have to predefine the $diff variable here
set_by_lua $sum '
local a = 32
local b = 56
ngx.var.diff = a - b; -- write to $diff directly
return a + b; -- return the $sum value normally
';
echo "sum = $sum, diff = $diff";
}
This directive can be freely mixed with all directives of the ngx_http_rewrite_module, set-misc-nginx-module, and array-var-nginx-module modules. All of these directives will run in the same order as they appear in the config file.
set $foo 32;
set_by_lua $bar 'return tonumber(ngx.var.foo) + 1';
set $baz "bar: $bar"; # $baz == "bar: 33"
As from the v0.5.0rc29
release, Nginx variable interpolation is disabled in the
argument of this directive and therefore, the dollar sign character ($
) can be used directly.
This directive requires the ngx_devel_kit module.
Back to TOC
syntax: set_by_lua_block $res { lua-script }
context: server, server if, location, location if
phase: rewrite
Similar to the set_by_lua directive except that
{}
) instead of in an NGINX string literal (which requires special character escaping), andFor example,
set_by_lua_block $res { return 32 + math.cos(32) }
# $res now has the value "32.834223360507" or alike.
No special escaping is required in the Lua code block.
This directive was first introduced in the v0.9.17
release.
Back to TOC
syntax: set_by_lua_file $res
context: server, server if, location, location if
phase: rewrite
Equivalent to set_by_lua, except that the file specified by
contains the Lua code, or, as from thev0.5.0rc32
release, the Lua/LuaJIT bytecode to be executed.
Nginx variable interpolation is supported in the
argument string of this directive. But special care must be taken for injection attacks.
When a relative path like foo/bar.lua
is given, they will be turned into the absolute path relative to the server prefix
path determined by the -p PATH
command-line option while starting the Nginx server.
When the Lua code cache is turned on (by default), the user code is loaded once at the first request and cached and the Nginx config must be reloaded each time the Lua source file is modified. The Lua code cache can be temporarily disabled during development by switching lua_code_cache off
in nginx.conf
to avoid reloading Nginx.
This directive requires the ngx_devel_kit module.
Back to TOC
syntax: content_by_lua
context: location, location if
phase: content
WARNING Since the v0.9.17
release, use of this directive is discouraged; use the new content_by_lua_block directive instead.
Acts as a "content handler" and executes Lua code string specified in
for every request. The Lua code may make API calls and is executed as a new spawned coroutine in an independent global environment (i.e. a sandbox).
Do not use this directive and other content handler directives in the same location. For example, this directive and the proxy_passdirective should not be used in the same location.
Back to TOC
syntax: content_by_lua_block { lua-script }
context: location, location if
phase: content
Similar to the content_by_lua directive except that this directive inlines the Lua source directly inside a pair of curly braces ({}
) instead of in an NGINX string literal (which requires special character escaping).
For instance,
content_by_lua_block {
ngx.say("I need no extra escaping here, for example: \r\nblah")
}
This directive was first introduced in the v0.9.17
release.
Back to TOC
syntax: content_by_lua_file
context: location, location if
phase: content
Equivalent to content_by_lua, except that the file specified by
contains the Lua code, or, as from the v0.5.0rc32
release, the Lua/LuaJIT bytecode to be executed.
Nginx variables can be used in the
string to provide flexibility. This however carries some risks and is not ordinarily recommended.
When a relative path like foo/bar.lua
is given, they will be turned into the absolute path relative to the server prefix
path determined by the -p PATH
command-line option while starting the Nginx server.
When the Lua code cache is turned on (by default), the user code is loaded once at the first request and cached and the Nginx config must be reloaded each time the Lua source file is modified. The Lua code cache can be temporarily disabled during development by switching lua_code_cache off
in nginx.conf
to avoid reloading Nginx.
Nginx variables are supported in the file path for dynamic dispatch, for example:
# WARNING: contents in nginx var must be carefully filtered,
# otherwise there'll be great security risk!
location ~ ^/app/([-_a-zA-Z0-9/]+) {
set $path $1;
content_by_lua_file /path/to/lua/app/root/$path.lua;
}
But be very careful about malicious user inputs and always carefully validate or filter out the user-supplied path components.
Back to TOC
syntax: rewrite_by_lua
context: http, server, location, location if
phase: rewrite tail
WARNING Since the v0.9.17
release, use of this directive is discouraged; use the new rewrite_by_lua_block directive instead.
Acts as a rewrite phase handler and executes Lua code string specified in
for every request. The Lua code may make API calls and is executed as a new spawned coroutine in an independent global environment (i.e. a sandbox).
Note that this handler always runs after the standard ngx_http_rewrite_module. So the following will work as expected:
location /foo {
set $a 12; # create and initialize $a
set $b ""; # create and initialize $b
rewrite_by_lua 'ngx.var.b = tonumber(ngx.var.a) + 1';
echo "res = $b";
}
because set $a 12
and set $b ""
run before rewrite_by_lua.
On the other hand, the following will not work as expected:
? location /foo {
? set $a 12; # create and initialize $a
? set $b ''; # create and initialize $b
? rewrite_by_lua 'ngx.var.b = tonumber(ngx.var.a) + 1';
? if ($b = '13') {
? rewrite ^ /bar redirect;
? break;
? }
?
? echo "res = $b";
? }
because if
runs before rewrite_by_lua even if it is placed after rewrite_by_lua in the config.
The right way of doing this is as follows:
location /foo {
set $a 12; # create and initialize $a
set $b ''; # create and initialize $b
rewrite_by_lua '
ngx.var.b = tonumber(ngx.var.a) + 1
if tonumber(ngx.var.b) == 13 then
return ngx.redirect("/bar");
end
';
echo "res = $b";
}
Note that the ngx_eval module can be approximated by using rewrite_by_lua. For example,
location / {
eval $res {
proxy_pass http://foo.com/check-spam;
}
if ($res = 'spam') {
rewrite ^ /terms-of-use.html redirect;
}
fastcgi_pass ...;
}
can be implemented in ngx_lua as:
location = /check-spam {
internal;
proxy_pass http://foo.com/check-spam;
}
location / {
rewrite_by_lua '
local res = ngx.location.capture("/check-spam")
if res.body == "spam" then
return ngx.redirect("/terms-of-use.html")
end
';
fastcgi_pass ...;
}
Just as any other rewrite phase handlers, rewrite_by_lua also runs in subrequests.
Note that when calling ngx.exit(ngx.OK)
within a rewrite_by_lua handler, the nginx request processing control flow will still continue to the content handler. To terminate the current request from within a rewrite_by_lua handler, calling ngx.exit with status >= 200 (ngx.HTTP_OK
) and status < 300 (ngx.HTTP_SPECIAL_RESPONSE
) for successful quits andngx.exit(ngx.HTTP_INTERNAL_SERVER_ERROR)
(or its friends) for failures.
If the ngx_http_rewrite_module's rewrite directive is used to change the URI and initiate location re-lookups (internal redirections), then any rewrite_by_lua or rewrite_by_lua_file code sequences within the current location will not be executed. For example,
location /foo {
rewrite ^ /bar;
rewrite_by_lua 'ngx.exit(503)';
}
location /bar {
...
}
Here the Lua code ngx.exit(503)
will never run. This will be the case if rewrite ^ /bar last
is used as this will similarly initiate an internal redirection. If the break
modifier is used instead, there will be no internal redirection and therewrite_by_lua
code will be executed.
The rewrite_by_lua
code will always run at the end of the rewrite
request-processing phase unlessrewrite_by_lua_no_postpone is turned on.
Back to TOC
syntax: rewrite_by_lua_block { lua-script }
context: http, server, location, location if
phase: rewrite tail
Similar to the rewrite_by_lua directive except that this directive inlines the Lua source directly inside a pair of curly braces ({}
) instead of in an NGINX string literal (which requires special character escaping).
For instance,
rewrite_by_lua_block {
do_something("hello, world!\nhiya\n")
}
This directive was first introduced in the v0.9.17
release.
Back to TOC
syntax: rewrite_by_lua_file
context: http, server, location, location if
phase: rewrite tail
Equivalent to rewrite_by_lua, except that the file specified by
contains the Lua code, or, as from the v0.5.0rc32
release, the Lua/LuaJIT bytecode to be executed.
Nginx variables can be used in the
string to provide flexibility. This however carries some risks and is not ordinarily recommended.
When a relative path like foo/bar.lua
is given, they will be turned into the absolute path relative to the server prefix
path determined by the -p PATH
command-line option while starting the Nginx server.
When the Lua code cache is turned on (by default), the user code is loaded once at the first request and cached and the Nginx config must be reloaded each time the Lua source file is modified. The Lua code cache can be temporarily disabled during development by switching lua_code_cache off
in nginx.conf
to avoid reloading Nginx.
The rewrite_by_lua_file
code will always run at the end of the rewrite
request-processing phase unlessrewrite_by_lua_no_postpone is turned on.
Nginx variables are supported in the file path for dynamic dispatch just as in content_by_lua_file.
Back to TOC
syntax: access_by_lua
context: http, server, location, location if
phase: access tail
WARNING Since the v0.9.17
release, use of this directive is discouraged; use the new access_by_lua_block directive instead.
Acts as an access phase handler and executes Lua code string specified in
for every request. The Lua code may make API calls and is executed as a new spawned coroutine in an independent global environment (i.e. a sandbox).
Note that this handler always runs after the standard ngx_http_access_module. So the following will work as expected:
location / {
deny 192.168.1.1;
allow 192.168.1.0/24;
allow 10.1.1.0/16;
deny all;
access_by_lua '
local res = ngx.location.capture("/mysql", { ... })
...
';
# proxy_pass/fastcgi_pass/...
}
That is, if a client IP address is in the blacklist, it will be denied before the MySQL query for more complex authentication is executed by access_by_lua.
Note that the ngx_auth_request module can be approximated by using access_by_lua:
location / {
auth_request /auth;
# proxy_pass/fastcgi_pass/postgres_pass/...
}
can be implemented in ngx_lua as:
location / {
access_by_lua '
local res = ngx.location.capture("/auth")
if res.status == ngx.HTTP_OK then
return
end
if res.status == ngx.HTTP_FORBIDDEN then
ngx.exit(res.status)
end
ngx.exit(ngx.HTTP_INTERNAL_SERVER_ERROR)
';
# proxy_pass/fastcgi_pass/postgres_pass/...
}
As with other access phase handlers, access_by_lua will not run in subrequests.
Note that when calling ngx.exit(ngx.OK)
within a access_by_lua handler, the nginx request processing control flow will still continue to the content handler. To terminate the current request from within a access_by_lua handler, calling ngx.exit with status >= 200 (ngx.HTTP_OK
) and status < 300 (ngx.HTTP_SPECIAL_RESPONSE
) for successful quits andngx.exit(ngx.HTTP_INTERNAL_SERVER_ERROR)
(or its friends) for failures.
Starting from the v0.9.20
release, you can use the access_by_lua_no_postpone directive to control when to run this handler inside the "access" request-processing phase of NGINX.
Back to TOC
syntax: access_by_lua_block { lua-script }
context: http, server, location, location if
phase: access tail
Similar to the access_by_lua directive except that this directive inlines the Lua source directly inside a pair of curly braces ({}
) instead of in an NGINX string literal (which requires special character escaping).
For instance,
access_by_lua_block {
do_something("hello, world!\nhiya\n")
}
This directive was first introduced in the v0.9.17
release.
Back to TOC
syntax: access_by_lua_file
context: http, server, location, location if
phase: access tail
Equivalent to access_by_lua, except that the file specified by
contains the Lua code, or, as from the v0.5.0rc32
release, the Lua/LuaJIT bytecode to be executed.
Nginx variables can be used in the
string to provide flexibility. This however carries some risks and is not ordinarily recommended.
When a relative path like foo/bar.lua
is given, they will be turned into the absolute path relative to the server prefix
path determined by the -p PATH
command-line option while starting the Nginx server.
When the Lua code cache is turned on (by default), the user code is loaded once at the first request and cached and the Nginx config must be reloaded each time the Lua source file is modified. The Lua code cache can be temporarily disabled during development by switching lua_code_cache off
in nginx.conf
to avoid repeatedly reloading Nginx.
Nginx variables are supported in the file path for dynamic dispatch just as in content_by_lua_file.
Back to TOC
syntax: header_filter_by_lua
context: http, server, location, location if
phase: output-header-filter
WARNING Since the v0.9.17
release, use of this directive is discouraged; use the new header_filter_by_lua_block directive instead.
Uses Lua code specified in
to define an output header filter.
Note that the following API functions are currently disabled within this context:
Here is an example of overriding a response header (or adding one if absent) in our Lua header filter:
location / {
proxy_pass http://mybackend;
header_filter_by_lua 'ngx.header.Foo = "blah"';
}
This directive was first introduced in the v0.2.1rc20
release.
Back to TOC
syntax: header_filter_by_lua_block { lua-script }
context: http, server, location, location if
phase: output-header-filter
Similar to the header_filter_by_lua directive except that this directive inlines the Lua source directly inside a pair of curly braces ({}
) instead of in an NGINX string literal (which requires special character escaping).
For instance,
header_filter_by_lua_block {
ngx.header["content-length"] = nil
}
This directive was first introduced in the v0.9.17
release.
Back to TOC
syntax: header_filter_by_lua_file
context: http, server, location, location if
phase: output-header-filter
Equivalent to header_filter_by_lua, except that the file specified by
contains the Lua code, or as from the v0.5.0rc32
release, the Lua/LuaJIT bytecode to be executed.
When a relative path like foo/bar.lua
is given, they will be turned into the absolute path relative to the server prefix
path determined by the -p PATH
command-line option while starting the Nginx server.
This directive was first introduced in the v0.2.1rc20
release.
Back to TOC
syntax: body_filter_by_lua
context: http, server, location, location if
phase: output-body-filter
WARNING Since the v0.9.17
release, use of this directive is discouraged; use the new body_filter_by_lua_block directive instead.
Uses Lua code specified in
to define an output body filter.
The input data chunk is passed via ngx.arg[1] (as a Lua string value) and the "eof" flag indicating the end of the response body data stream is passed via ngx.arg[2] (as a Lua boolean value).
Behind the scene, the "eof" flag is just the last_buf
(for main requests) or last_in_chain
(for subrequests) flag of the Nginx chain link buffers. (Before the v0.7.14
release, the "eof" flag does not work at all in subrequests.)
The output data stream can be aborted immediately by running the following Lua statement:
return ngx.ERROR
This will truncate the response body and usually result in incomplete and also invalid responses.
The Lua code can pass its own modified version of the input data chunk to the downstream Nginx output body filters by overridingngx.arg[1] with a Lua string or a Lua table of strings. For example, to transform all the lowercase letters in the response body, we can just write:
location / {
proxy_pass http://mybackend;
body_filter_by_lua 'ngx.arg[1] = string.upper(ngx.arg[1])';
}
When setting nil
or an empty Lua string value to ngx.arg[1]
, no data chunk will be passed to the downstream Nginx output filters at all.
Likewise, new "eof" flag can also be specified by setting a boolean value to ngx.arg[2]. For example,
location /t {
echo hello world;
echo hiya globe;
body_filter_by_lua '
local chunk = ngx.arg[1]
if string.match(chunk, "hello") then
ngx.arg[2] = true -- new eof
return
end
-- just throw away any remaining chunk data
ngx.arg[1] = nil
';
}
Then GET /t
will just return the output
hello world
That is, when the body filter sees a chunk containing the word "hello", then it will set the "eof" flag to true immediately, resulting in truncated but still valid responses.
When the Lua code may change the length of the response body, then it is required to always clear out the Content-Length
response header (if any) in a header filter to enforce streaming output, as in
location /foo {
# fastcgi_pass/proxy_pass/...
header_filter_by_lua 'ngx.header.content_length = nil';
body_filter_by_lua 'ngx.arg[1] = string.len(ngx.arg[1]) .. "\\n"';
}
Note that the following API functions are currently disabled within this context due to the limitations in NGINX output filter's current implementation:
Nginx output filters may be called multiple times for a single request because response body may be delivered in chunks. Thus, the Lua code specified by in this directive may also run multiple times in the lifetime of a single HTTP request.
This directive was first introduced in the v0.5.0rc32
release.
Back to TOC
syntax: body_filter_by_lua_block { lua-script-str }
context: http, server, location, location if
phase: output-body-filter
Similar to the body_filter_by_lua directive except that this directive inlines the Lua source directly inside a pair of curly braces ({}
) instead of in an NGINX string literal (which requires special character escaping).
For instance,
body_filter_by_lua_block {
local data, eof = ngx.arg[1], ngx.arg[2]
}
This directive was first introduced in the v0.9.17
release.
Back to TOC
syntax: body_filter_by_lua_file
context: http, server, location, location if
phase: output-body-filter
Equivalent to body_filter_by_lua, except that the file specified by
contains the Lua code, or, as from the v0.5.0rc32
release, the Lua/LuaJIT bytecode to be executed.
When a relative path like foo/bar.lua
is given, they will be turned into the absolute path relative to the server prefix
path determined by the -p PATH
command-line option while starting the Nginx server.
This directive was first introduced in the v0.5.0rc32
release.
Back to TOC
syntax: log_by_lua
context: http, server, location, location if
phase: log
WARNING Since the v0.9.17
release, use of this directive is discouraged; use the new log_by_lua_block directive instead.
Runs the Lua source code inlined as the
at the log
request processing phase. This does not replace the current access logs, but runs before.
Note that the following API functions are currently disabled within this context:
Here is an example of gathering average data for $upstream_response_time:
lua_shared_dict log_dict 5M;
server {
location / {
proxy_pass http://mybackend;
log_by_lua '
local log_dict = ngx.shared.log_dict
local upstream_time = tonumber(ngx.var.upstream_response_time)
local sum = log_dict:get("upstream_time-sum") or 0
sum = sum + upstream_time
log_dict:set("upstream_time-sum", sum)
local newval, err = log_dict:incr("upstream_time-nb", 1)
if not newval and err == "not found" then
log_dict:add("upstream_time-nb", 0)
log_dict:incr("upstream_time-nb", 1)
end
';
}
location = /status {
content_by_lua '
local log_dict = ngx.shared.log_dict
local sum = log_dict:get("upstream_time-sum")
local nb = log_dict:get("upstream_time-nb")
if nb and sum then
ngx.say("average upstream response time: ", sum / nb,
" (", nb, " reqs)")
else
ngx.say("no data yet")
end
';
}
}
This directive was first introduced in the v0.5.0rc31
release.
Back to TOC
syntax: log_by_lua_block { lua-script }
context: http, server, location, location if
phase: log
Similar to the log_by_lua directive except that this directive inlines the Lua source directly inside a pair of curly braces ({}
) instead of in an NGINX string literal (which requires special character escaping).
For instance,
log_by_lua_block {
print("I need no extra escaping here, for example: \r\nblah")
}
This directive was first introduced in the v0.9.17
release.
Back to TOC
syntax: log_by_lua_file
context: http, server, location, location if
phase: log
Equivalent to log_by_lua, except that the file specified by
contains the Lua code, or, as from thev0.5.0rc32
release, the Lua/LuaJIT bytecode to be executed.
When a relative path like foo/bar.lua
is given, they will be turned into the absolute path relative to the server prefix
path determined by the -p PATH
command-line option while starting the Nginx server.
This directive was first introduced in the v0.5.0rc31
release.
Back to TOC
syntax: balancer_by_lua_block { lua-script }
context: upstream
phase: content
This directive runs Lua code as an upstream balancer for any upstream entities defined by the upstream {}
configuration block.
For instance,
upstream foo {
server 127.0.0.1;
balancer_by_lua_block {
-- use Lua to do something interesting here
-- as a dynamic balancer
}
}
server {
location / {
proxy_pass http://foo;
}
}
The resulting Lua load balancer can work with any existing nginx upstream modules like ngx_proxy and ngx_fastcgi.
Also, the Lua load balancer can work with the standard upstream connection pool mechanism, i.e., the standard keepalivedirective. Just ensure that the keepalive directive is used after this balancer_by_lua_block
directive in a single upstream {}
configuration block.
The Lua load balancer can totally ignore the list of servers defined in the upstream {}
block and select peer from a completely dynamic server list (even changing per request) via the ngx.balancer module from the lua-resty-core library.
The Lua code handler registered by this directive might get called more than once in a single downstream request when the nginx upstream mechanism retries the request on conditions specified by directives like the proxy_next_upstream directive.
This Lua code execution context does not support yielding, so Lua APIs that may yield (like cosockets and "light threads") are disabled in this context. One can usually work around this limitation by doing such operations in an earlier phase handler (likeaccess_by_lua*) and passing along the result into this context via the ngx.ctx table.
This directive was first introduced in the v0.10.0
release.
Back to TOC
syntax: balancer_by_lua_file
context: upstream
phase: content
Equivalent to balancer_by_lua_block, except that the file specified by
contains the Lua code, or, as from the v0.5.0rc32
release, the Lua/LuaJIT bytecode to be executed.
When a relative path like foo/bar.lua
is given, they will be turned into the absolute path relative to the server prefix
path determined by the -p PATH
command-line option while starting the Nginx server.
This directive was first introduced in the v0.10.0
release.
Back to TOC
syntax: lua_need_request_body
default: off
context: http, server, location, location if
phase: depends on usage
Determines whether to force the request body data to be read before running rewrite/access/access_by_lua* or not. The Nginx core does not read the client request body by default and if request body data is required, then this directive should be turned on
or the ngx.req.read_body function should be called within the Lua code.
To read the request body data within the $request_body variable, client_body_buffer_size must have the same value asclient_max_body_size. Because when the content length exceeds client_body_buffer_size but less than client_max_body_size, Nginx will buffer the data into a temporary file on the disk, which will lead to empty value in the $request_body variable.
If the current location includes rewrite_by_lua or rewrite_by_lua_file directives, then the request body will be read just before therewrite_by_lua or rewrite_by_lua_file code is run (and also at the rewrite
phase). Similarly, if only content_by_lua is specified, the request body will not be read until the content handler's Lua code is about to run (i.e., the request body will be read during the content phase).
It is recommended however, to use the ngx.req.read_body and ngx.req.discard_body functions for finer control over the request body reading process instead.
This also applies to access_by_lua and access_by_lua_file.
Back to TOC
syntax: ssl_certificate_by_lua_block { lua-script }
context: server
phase: right-before-SSL-handshake
This directive runs user Lua code when NGINX is about to start the SSL handshake for the downstream SSL (https) connections.
It is particularly useful for setting the SSL certificate chain and the corresponding private key on a per-request basis. It is also useful to load such handshake configurations nonblockingly from the remote (for example, with the cosocket API). And one can also do per-request OCSP stapling handling in pure Lua here as well.
Another typical use case is to do SSL handshake traffic control nonblockingly in this context, with the help of the lua-resty-limit-traffic#readme library, for example.
One can also do interesting things with the SSL handshake requests from the client side, like rejecting old SSL clients using the SSLv3 protocol or even below selectively.
The ngx.ssl and ngx.ocsp Lua modules provided by the lua-resty-core library are particularly useful in this context. You can use the Lua API offered by these two Lua modules to manipulate the SSL certificate chain and private key for the current SSL connection being initiated.
This Lua handler does not run at all, however, when NGINX/OpenSSL successfully resumes the SSL session via SSL session IDs or TLS session tickets for the current SSL connection. In other words, this Lua handler only runs when NGINX has to initiate a full SSL handshake.
Below is a trivial example using the ngx.ssl module at the same time:
server {
listen 443 ssl;
server_name test.com;
ssl_certificate_by_lua_block {
print("About to initiate a new SSL handshake!")
}
location / {
root html;
}
}
See more complicated examples in the ngx.ssl and ngx.ocsp Lua modules' official documentation.
Uncaught Lua exceptions in the user Lua code immediately abort the current SSL session, so does the ngx.exit call with an error code like ngx.ERROR
.
This Lua code execution context does support yielding, so Lua APIs that may yield (like cosockets, sleeping, and "light threads") are enabled in this context.
Note, however, you still need to configure the ssl_certificate and ssl_certificate_key directives even though you will not use this static certificate and private key at all. This is because the NGINX core requires their appearance otherwise you are seeing the following error while starting NGINX:
nginx: [emerg] no ssl configured for the server
This directive currently requires the following NGINX core patch to work correctly:
http://mailman.nginx.org/pipermail/nginx-devel/2016-January/007748.html
The bundled version of the NGINX core in OpenResty 1.9.7.2 (or above) already has this patch applied.
Furthermore, one needs at least OpenSSL 1.0.2e for this directive to work.
This directive was first introduced in the v0.10.0
release.
Back to TOC
syntax: ssl_certificate_by_lua_file
context: server
phase: right-before-SSL-handshake
Equivalent to ssl_certificate_by_lua_block, except that the file specified by
contains the Lua code, or, as from the v0.5.0rc32
release, the Lua/LuaJIT bytecode to be executed.
When a relative path like foo/bar.lua
is given, they will be turned into the absolute path relative to the server prefix
path determined by the -p PATH
command-line option while starting the Nginx server.
This directive was first introduced in the v0.10.0
release.
Back to TOC
syntax: lua_shared_dict
default: no
context: http
phase: depends on usage
Declares a shared memory zone,
, to serve as storage for the shm based Lua dictionary ngx.shared.
.
Shared memory zones are always shared by all the nginx worker processes in the current nginx server instance.
The
argument accepts size units such as k
and m
:
http {
lua_shared_dict dogs 10m;
...
}
The hard-coded minimum size is 8KB while the practical minimum size depends on actual user data set (some people start with 12KB).
See ngx.shared.DICT for details.
This directive was first introduced in the v0.3.1rc22
release.
Back to TOC
syntax: lua_socket_connect_timeout
default: lua_socket_connect_timeout 60s
context: http, server, location
This directive controls the default timeout value used in TCP/unix-domain socket object's connect method and can be overridden by the settimeout method.
The argument can be an integer, with an optional time unit, like
s
(second), ms
(millisecond), m
(minute). The default time unit is s
, i.e., "second". The default setting is 60s
.
This directive was first introduced in the v0.5.0rc1
release.
Back to TOC
syntax: lua_socket_send_timeout
default: lua_socket_send_timeout 60s
context: http, server, location
Controls the default timeout value used in TCP/unix-domain socket object's send method and can be overridden by the settimeoutmethod.
The argument can be an integer, with an optional time unit, like
s
(second), ms
(millisecond), m
(minute). The default time unit is s
, i.e., "second". The default setting is 60s
.
This directive was first introduced in the v0.5.0rc1
release.
Back to TOC
syntax: lua_socket_send_lowat
default: lua_socket_send_lowat 0
context: http, server, location
Controls the lowat
(low water) value for the cosocket send buffer.
Back to TOC
syntax: lua_socket_read_timeout
default: lua_socket_read_timeout 60s
context: http, server, location
phase: depends on usage
This directive controls the default timeout value used in TCP/unix-domain socket object's receive method and iterator functions returned by the receiveuntil method. This setting can be overridden by the settimeout method.
The argument can be an integer, with an optional time unit, like
s
(second), ms
(millisecond), m
(minute). The default time unit is s
, i.e., "second". The default setting is 60s
.
This directive was first introduced in the v0.5.0rc1
release.
Back to TOC
syntax: lua_socket_buffer_size
default: lua_socket_buffer_size 4k/8k
context: http, server, location
Specifies the buffer size used by cosocket reading operations.
This buffer does not have to be that big to hold everything at the same time because cosocket supports 100% non-buffered reading and parsing. So even 1
byte buffer size should still work everywhere but the performance could be terrible.
This directive was first introduced in the v0.5.0rc1
release.
Back to TOC
syntax: lua_socket_pool_size
default: lua_socket_pool_size 30
context: http, server, location
Specifies the size limit (in terms of connection count) for every cosocket connection pool associated with every remote server (i.e., identified by either the host-port pair or the unix domain socket file path).
Default to 30 connections for every pool.
When the connection pool exceeds the available size limit, the least recently used (idle) connection already in the pool will be closed to make room for the current connection.
Note that the cosocket connection pool is per nginx worker process rather than per nginx server instance, so size limit specified here also applies to every single nginx worker process.
This directive was first introduced in the v0.5.0rc1
release.
Back to TOC
syntax: lua_socket_keepalive_timeout
default: lua_socket_keepalive_timeout 60s
context: http, server, location
This directive controls the default maximal idle time of the connections in the cosocket built-in connection pool. When this timeout reaches, idle connections will be closed and removed from the pool. This setting can be overridden by cosocket objects'setkeepalive method.
The argument can be an integer, with an optional time unit, like
s
(second), ms
(millisecond), m
(minute). The default time unit is s
, i.e., "second". The default setting is 60s
.
This directive was first introduced in the v0.5.0rc1
release.
Back to TOC
syntax: lua_socket_log_errors on|off
default: lua_socket_log_errors on
context: http, server, location
This directive can be used to toggle error logging when a failure occurs for the TCP or UDP cosockets. If you are already doing proper error handling and logging in your Lua code, then it is recommended to turn this directive off to prevent data flushing in your nginx error log files (which is usually rather expensive).
This directive was first introduced in the v0.5.13
release.
Back to TOC
syntax: lua_ssl_ciphers
default: lua_ssl_ciphers DEFAULT
context: http, server, location
Specifies the enabled ciphers for requests to a SSL/TLS server in the tcpsock:sslhandshake method. The ciphers are specified in the format understood by the OpenSSL library.
The full list can be viewed using the “openssl ciphers” command.
This directive was first introduced in the v0.9.11
release.
Back to TOC
syntax: lua_ssl_crl
default: no
context: http, server, location
Specifies a file with revoked certificates (CRL) in the PEM format used to verify the certificate of the SSL/TLS server in thetcpsock:sslhandshake method.
This directive was first introduced in the v0.9.11
release.
Back to TOC
syntax: lua_ssl_protocols [SSLv2] [SSLv3] [TLSv1] [TLSv1.1] [TLSv1.2]
default: lua_ssl_protocols SSLv3 TLSv1 TLSv1.1 TLSv1.2
context: http, server, location
Enables the specified protocols for requests to a SSL/TLS server in the tcpsock:sslhandshake method.
This directive was first introduced in the v0.9.11
release.
Back to TOC
syntax: lua_ssl_trusted_certificate
default: no
context: http, server, location
Specifies a file path with trusted CA certificates in the PEM format used to verify the certificate of the SSL/TLS server in thetcpsock:sslhandshake method.
This directive was first introduced in the v0.9.11
release.
See also lua_ssl_verify_depth.
Back to TOC
syntax: lua_ssl_verify_depth
default: lua_ssl_verify_depth 1
context: http, server, location
Sets the verification depth in the server certificates chain.
This directive was first introduced in the v0.9.11
release.
See also lua_ssl_trusted_certificate.
Back to TOC
syntax: lua_http10_buffering on|off
default: lua_http10_buffering on
context: http, server, location, location-if
Enables or disables automatic response buffering for HTTP 1.0 (or older) requests. This buffering mechanism is mainly used for HTTP 1.0 keep-alive which replies on a proper Content-Length
response header.
If the Lua code explicitly sets a Content-Length
response header before sending the headers (either explicitly viangx.send_headers or implicitly via the first ngx.say or ngx.print call), then the HTTP 1.0 response buffering will be disabled even when this directive is turned on.
To output very large response data in a streaming fashion (via the ngx.flush call, for example), this directive MUST be turned off to minimize memory usage.
This directive is turned on
by default.
This directive was first introduced in the v0.5.0rc19
release.
Back to TOC
syntax: rewrite_by_lua_no_postpone on|off
default: rewrite_by_lua_no_postpone off
context: http
Controls whether or not to disable postponing rewrite_by_lua* directives to run at the end of the rewrite
request-processing phase. By default, this directive is turned off and the Lua code is postponed to run at the end of the rewrite
phase.
This directive was first introduced in the v0.5.0rc29
release.
Back to TOC
syntax: access_by_lua_no_postpone on|off
default: access_by_lua_no_postpone off
context: http
Controls whether or not to disable postponing access_by_lua* directives to run at the end of the access
request-processing phase. By default, this directive is turned off and the Lua code is postponed to run at the end of the access
phase.
This directive was first introduced in the v0.9.20
release.
Back to TOC
syntax: lua_transform_underscores_in_response_headers on|off
default: lua_transform_underscores_in_response_headers on
context: http, server, location, location-if
Controls whether to transform underscores (_
) in the response header names specified in the ngx.header.HEADER API to hypens (-
).
This directive was first introduced in the v0.5.0rc32
release.
Back to TOC
syntax: lua_check_client_abort on|off
default: lua_check_client_abort off
context: http, server, location, location-if
This directive controls whether to check for premature client connection abortion.
When this directive is turned on, the ngx_lua module will monitor the premature connection close event on the downstream connections. And when there is such an event, it will call the user Lua function callback (registered by ngx.on_abort) or just stop and clean up all the Lua "light threads" running in the current request's request handler when there is no user callback function registered.
According to the current implementation, however, if the client closes the connection before the Lua code finishes reading the request body data via ngx.req.socket, then ngx_lua will neither stop all the running "light threads" nor call the user callback (ifngx.on_abort has been called). Instead, the reading operation on ngx.req.socket will just return the error message "client aborted" as the second return value (the first return value is surely nil
).
When TCP keepalive is disabled, it is relying on the client side to close the socket gracefully (by sending a FIN
packet or something like that). For (soft) real-time web applications, it is highly recommended to configure the TCP keepalive support in your system's TCP stack implementation in order to detect "half-open" TCP connections in time.
For example, on Linux, you can configure the standard listen directive in your nginx.conf
file like this:
listen 80 so_keepalive=2s:2s:8;
On FreeBSD, you can only tune the system-wide configuration for TCP keepalive, for example:
# sysctl net.inet.tcp.keepintvl=2000
# sysctl net.inet.tcp.keepidle=2000
This directive was first introduced in the v0.7.4
release.
See also ngx.on_abort.
Back to TOC
syntax: lua_max_pending_timers
default: lua_max_pending_timers 1024
context: http
Controls the maximum number of pending timers allowed.
Pending timers are those timers that have not expired yet.
When exceeding this limit, the ngx.timer.at call will immediately return nil
and the error string "too many pending timers".
This directive was first introduced in the v0.8.0
release.
Back to TOC
syntax: lua_max_running_timers
default: lua_max_running_timers 256
context: http
Controls the maximum number of "running timers" allowed.
Running timers are those timers whose user callback functions are still running.
When exceeding this limit, Nginx will stop running the callbacks of newly expired timers and log an error message "N lua_max_running_timers are not enough" where "N" is the current value of this directive.
This directive was first introduced in the v0.8.0
release.
Back to TOC
Back to TOC
The various *_by_lua
and *_by_lua_file
configuration directives serve as gateways to the Lua API within the nginx.conf
file. The Nginx Lua API described below can only be called within the user Lua code run in the context of these configuration directives.
The API is exposed to Lua in the form of two standard packages ngx
and ndk
. These packages are in the default global scope within ngx_lua and are always available within ngx_lua directives.
The packages can be introduced into external Lua modules like this:
local say = ngx.say
local _M = {}
function _M.foo(a)
say(a)
end
return _M
Use of the package.seeall flag is strongly discouraged due to its various bad side-effects.
It is also possible to directly require the packages in external Lua modules:
local ngx = require "ngx"
local ndk = require "ndk"
The ability to require these packages was introduced in the v0.2.1rc19
release.
Network I/O operations in user code should only be done through the Nginx Lua API calls as the Nginx event loop may be blocked and performance drop off dramatically otherwise. Disk operations with relatively small amount of data can be done using the standard Lua io
library but huge file reading and writing should be avoided wherever possible as they may block the Nginx process significantly. Delegating all network and disk I/O operations to Nginx's subrequests (via the ngx.location.capture method and similar) is strongly recommended for maximum performance.
Back to TOC
syntax: val = ngx.arg[index]
context: set_by_lua*, body_filter_by_lua*
When this is used in the context of the set_by_lua or set_by_lua_file directives, this table is read-only and holds the input arguments to the config directives:
value = ngx.arg[n]
Here is an example
location /foo {
set $a 32;
set $b 56;
set_by_lua $sum
'return tonumber(ngx.arg[1]) + tonumber(ngx.arg[2])'
$a $b;
echo $sum;
}
that writes out 88
, the sum of 32
and 56
.
When this table is used in the context of body_filter_by_lua or body_filter_by_lua_file, the first element holds the input data chunk to the output filter code and the second element holds the boolean flag for the "eof" flag indicating the end of the whole output data stream.
The data chunk and "eof" flag passed to the downstream Nginx output filters can also be overridden by assigning values directly to the corresponding table elements. When setting nil
or an empty Lua string value to ngx.arg[1]
, no data chunk will be passed to the downstream Nginx output filters at all.
Back to TOC
syntax: ngx.var.VAR_NAME
context: set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, body_filter_by_lua*, log_by_lua*
Read and write Nginx variable values.
value = ngx.var.some_nginx_variable_name
ngx.var.some_nginx_variable_name = value
Note that only already defined nginx variables can be written to. For example:
location /foo {
set $my_var ''; # this line is required to create $my_var at config time
content_by_lua '
ngx.var.my_var = 123;
...
';
}
That is, nginx variables cannot be created on-the-fly.
Some special nginx variables like $args
and $limit_rate
can be assigned a value, many others are not, like$query_string
, $arg_PARAMETER
, and $http_NAME
.
Nginx regex group capturing variables $1
, $2
, $3
, and etc, can be read by this interface as well, by writing ngx.var[1]
,ngx.var[2]
, ngx.var[3]
, and etc.
Setting ngx.var.Foo
to a nil
value will unset the $Foo
Nginx variable.
ngx.var.args = nil
WARNING When reading from an Nginx variable, Nginx will allocate memory in the per-request memory pool which is freed only at request termination. So when you need to read from an Nginx variable repeatedly in your Lua code, cache the Nginx variable value to your own Lua variable, for example,
local val = ngx.var.some_var
--- use the val repeatedly later
to prevent (temporary) memory leaking within the current request's lifetime. Another way of caching the result is to use the ngx.ctxtable.
Undefined NGINX variables are evaluated to nil
while uninitialized (but defined) NGINX variables are evaluated to an empty Lua string.
This API requires a relatively expensive metamethod call and it is recommended to avoid using it on hot code paths.
Back to TOC
context: init_by_lua*, set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, body_filter_by_lua, *log_by_lua*, ngx.timer.*, balancer_by_lua*, ssl_certificate_by_lua*
ngx.OK (0)
ngx.ERROR (-1)
ngx.AGAIN (-2)
ngx.DONE (-4)
ngx.DECLINED (-5)
Note that only three of these constants are utilized by the Nginx API for Lua (i.e., ngx.exit accepts NGX_OK
, NGX_ERROR
, andNGX_DECLINED
as input).
ngx.null
The ngx.null
constant is a NULL
light userdata usually used to represent nil values in Lua tables etc and is similar to the lua-cjson library's cjson.null
constant. This constant was first introduced in the v0.5.0rc5
release.
The ngx.DECLINED
constant was first introduced in the v0.5.0rc19
release.
Back to TOC
context: init_by_lua*, set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, body_filter_by_lua, log_by_lua*, ngx.timer.*, balancer_by_lua*, ssl_certificate_by_lua*
ngx.HTTP_GET
ngx.HTTP_HEAD
ngx.HTTP_PUT
ngx.HTTP_POST
ngx.HTTP_DELETE
ngx.HTTP_OPTIONS (added in the v0.5.0rc24 release)
ngx.HTTP_MKCOL (added in the v0.8.2 release)
ngx.HTTP_COPY (added in the v0.8.2 release)
ngx.HTTP_MOVE (added in the v0.8.2 release)
ngx.HTTP_PROPFIND (added in the v0.8.2 release)
ngx.HTTP_PROPPATCH (added in the v0.8.2 release)
ngx.HTTP_LOCK (added in the v0.8.2 release)
ngx.HTTP_UNLOCK (added in the v0.8.2 release)
ngx.HTTP_PATCH (added in the v0.8.2 release)
ngx.HTTP_TRACE (added in the v0.8.2 release)
These constants are usually used in ngx.location.capture and ngx.location.capture_multi method calls.
Back to TOC
context: init_by_lua*, set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, body_filter_by_lua, log_by_lua*, ngx.timer.*, balancer_by_lua*, ssl_certificate_by_lua*
value = ngx.HTTP_CONTINUE (100) (first added in the v0.9.20 release)
value = ngx.HTTP_SWITCHING_PROTOCOLS (101) (first added in the v0.9.20 release)
value = ngx.HTTP_OK (200)
value = ngx.HTTP_CREATED (201)
value = ngx.HTTP_ACCEPTED (202) (first added in the v0.9.20 release)
value = ngx.HTTP_NO_CONTENT (204) (first added in the v0.9.20 release)
value = ngx.HTTP_PARTIAL_CONTENT (206) (first added in the v0.9.20 release)
value = ngx.HTTP_SPECIAL_RESPONSE (300)
value = ngx.HTTP_MOVED_PERMANENTLY (301)
value = ngx.HTTP_MOVED_TEMPORARILY (302)
value = ngx.HTTP_SEE_OTHER (303)
value = ngx.HTTP_NOT_MODIFIED (304)
value = ngx.HTTP_TEMPORARY_REDIRECT (307) (first added in the v0.9.20 release)
value = ngx.HTTP_BAD_REQUEST (400)
value = ngx.HTTP_UNAUTHORIZED (401)
value = ngx.HTTP_PAYMENT_REQUIRED (402) (first added in the v0.9.20 release)
value = ngx.HTTP_FORBIDDEN (403)
value = ngx.HTTP_NOT_FOUND (404)
value = ngx.HTTP_NOT_ALLOWED (405)
value = ngx.HTTP_NOT_ACCEPTABLE (406) (first added in the v0.9.20 release)
value = ngx.HTTP_REQUEST_TIMEOUT (408) (first added in the v0.9.20 release)
value = ngx.HTTP_CONFLICT (409) (first added in the v0.9.20 release)
value = ngx.HTTP_GONE (410)
value = ngx.HTTP_UPGRADE_REQUIRED (426) (first added in the v0.9.20 release)
value = ngx.HTTP_TOO_MANY_REQUESTS (429) (first added in the v0.9.20 release)
value = ngx.HTTP_CLOSE (444) (first added in the v0.9.20 release)
value = ngx.HTTP_ILLEGAL (451) (first added in the v0.9.20 release)
value = ngx.HTTP_INTERNAL_SERVER_ERROR (500)
value = ngx.HTTP_METHOD_NOT_IMPLEMENTED (501)
value = ngx.HTTP_BAD_GATEWAY (502) (first added in the v0.9.20 release)
value = ngx.HTTP_SERVICE_UNAVAILABLE (503)
value = ngx.HTTP_GATEWAY_TIMEOUT (504) (first added in the v0.3.1rc38 release)
value = ngx.HTTP_VERSION_NOT_SUPPORTED (505) (first added in the v0.9.20 release)
value = ngx.HTTP_INSUFFICIENT_STORAGE (507) (first added in the v0.9.20 release)
Back to TOC
context: set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, body_filter_by_lua, log_by_lua*, ngx.timer.*, balancer_by_lua*, ssl_certificate_by_lua*
ngx.STDERR
ngx.EMERG
ngx.ALERT
ngx.CRIT
ngx.ERR
ngx.WARN
ngx.NOTICE
ngx.INFO
ngx.DEBUG
These constants are usually used by the ngx.log method.
Back to TOC
syntax: print(...)
context: init_by_lua*, init_worker_by_lua*, set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, body_filter_by_lua, log_by_lua*, ngx.timer.*, balancer_by_lua*, certificate_by_lua*
Writes argument values into the nginx error.log
file with the ngx.NOTICE
log level.
It is equivalent to
ngx.log(ngx.NOTICE, ...)
Lua nil
arguments are accepted and result in literal "nil"
strings while Lua booleans result in literal "true"
or "false"
strings. And the ngx.null
constant will yield the "null"
string output.
There is a hard coded 2048
byte limitation on error message lengths in the Nginx core. This limit includes trailing newlines and leading time stamps. If the message size exceeds this limit, Nginx will truncate the message text accordingly. This limit can be manually modified by editing the NGX_MAX_ERROR_STR
macro definition in the src/core/ngx_log.h
file in the Nginx source tree.
Back to TOC
context: init_worker_by_lua*, set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, body_filter_by_lua, log_by_lua*, ngx.timer.*, balancer_by_lua*
This table can be used to store per-request Lua context data and has a life time identical to the current request (as with the Nginx variables).
Consider the following example,
location /test {
rewrite_by_lua '
ngx.ctx.foo = 76
';
access_by_lua '
ngx.ctx.foo = ngx.ctx.foo + 3
';
content_by_lua '
ngx.say(ngx.ctx.foo)
';
}
Then GET /test
will yield the output
79
That is, the ngx.ctx.foo
entry persists across the rewrite, access, and content phases of a request.
Every request, including subrequests, has its own copy of the table. For example:
location /sub {
content_by_lua '
ngx.say("sub pre: ", ngx.ctx.blah)
ngx.ctx.blah = 32
ngx.say("sub post: ", ngx.ctx.blah)
';
}
location /main {
content_by_lua '
ngx.ctx.blah = 73
ngx.say("main pre: ", ngx.ctx.blah)
local res = ngx.location.capture("/sub")
ngx.print(res.body)
ngx.say("main post: ", ngx.ctx.blah)
';
}
Then GET /main
will give the output
main pre: 73
sub pre: nil
sub post: 32
main post: 73
Here, modification of the ngx.ctx.blah
entry in the subrequest does not affect the one in the parent request. This is because they have two separate versions of ngx.ctx.blah
.
Internal redirection will destroy the original request ngx.ctx
data (if any) and the new request will have an empty ngx.ctx
table. For instance,
location /new {
content_by_lua '
ngx.say(ngx.ctx.foo)
';
}
location /orig {
content_by_lua '
ngx.ctx.foo = "hello"
ngx.exec("/new")
';
}
Then GET /orig
will give
nil
rather than the original "hello"
value.
Arbitrary data values, including Lua closures and nested tables, can be inserted into this "magic" table. It also allows the registration of custom meta methods.
Overriding ngx.ctx
with a new Lua table is also supported, for example,
ngx.ctx = { foo = 32, bar = 54 }
When being used in the context of init_worker_by_lua*, this table just has the same lifetime of the current Lua handler.
The ngx.ctx
lookup requires relatively expensive metamethod calls and it is much slower than explicitly passing per-request data along by your own function arguments. So do not abuse this API for saving your own function arguments because it usually has quite some performance impact.
Because of the metamethod magic, never "local" the ngx.ctx
table outside your Lua function scope on the Lua module level level due to worker-level data sharing. For example, the following is bad:
-- mymodule.lua
local _M = {}
-- the following line is bad since ngx.ctx is a per-request
-- data while this `ctx` variable is on the Lua module level
-- and thus is per-nginx-worker.
local ctx = ngx.ctx
function _M.main()
ctx.foo = "bar"
end
return _M
Use the following instead:
-- mymodule.lua
local _M = {}
function _M.main(ctx)
ctx.foo = "bar"
end
return _M
That is, let the caller pass the ctx
table explicitly via a function argument.
Back to TOC
syntax: res = ngx.location.capture(uri, options?)
context: rewrite_by_lua*, access_by_lua*, content_by_lua*
Issues a synchronous but still non-blocking Nginx Subrequest using uri
.
Nginx's subrequests provide a powerful way to make non-blocking internal requests to other locations configured with disk file directory or any other nginx C modules like ngx_proxy
, ngx_fastcgi
, ngx_memc
, ngx_postgres
, ngx_drizzle
, and even ngx_lua itself and etc etc etc.
Also note that subrequests just mimic the HTTP interface but there is no extra HTTP/TCP traffic nor IPC involved. Everything works internally, efficiently, on the C level.
Subrequests are completely different from HTTP 301/302 redirection (via ngx.redirect) and internal redirection (via ngx.exec).
You should always read the request body (by either calling ngx.req.read_body or configuring lua_need_request_body on) before initiating a subrequest.
This API function (as well as ngx.location.capture_multi) always buffers the whole response body of the subrequest in memory. Thus, you should use cosockets and streaming processing instead if you have to handle large subrequest responses.
Here is a basic example:
res = ngx.location.capture(uri)
Returns a Lua table with 4 slots: res.status
, res.header
, res.body
, and res.truncated
.
res.status
holds the response status code for the subrequest response.
res.header
holds all the response headers of the subrequest and it is a normal Lua table. For multi-value response headers, the value is a Lua (array) table that holds all the values in the order that they appear. For instance, if the subrequest response headers contain the following lines:
Set-Cookie: a=3
Set-Cookie: foo=bar
Set-Cookie: baz=blah
Then res.header["Set-Cookie"]
will be evaluated to the table value {"a=3", "foo=bar", "baz=blah"}
.
res.body
holds the subrequest's response body data, which might be truncated. You always need to check the res.truncated
boolean flag to see if res.body
contains truncated data. The data truncation here can only be caused by those unrecoverable errors in your subrequests like the cases that the remote end aborts the connection prematurely in the middle of the response body data stream or a read timeout happens when your subrequest is receiving the response body data from the remote.
URI query strings can be concatenated to URI itself, for instance,
res = ngx.location.capture('/foo/bar?a=3&b=4')
Named locations like @foo
are not allowed due to a limitation in the nginx core. Use normal locations combined with theinternal
directive to prepare internal-only locations.
An optional option table can be fed as the second argument, which supports the options:
method
specify the subrequest's request method, which only accepts constants like ngx.HTTP_POST
.body
specify the subrequest's request body (string value only).args
specify the subrequest's URI query arguments (both string value and Lua tables are accepted)ctx
specify a Lua table to be the ngx.ctx table for the subrequest. It can be the current request's ngx.ctx table, which effectively makes the parent and its subrequest to share exactly the same context table. This option was first introduced in thev0.3.1rc25
release.vars
take a Lua table which holds the values to set the specified Nginx variables in the subrequest as this option's value. This option was first introduced in the v0.3.1rc31
release.copy_all_vars
specify whether to copy over all the Nginx variable values of the current request to the subrequest in question. modifications of the nginx variables in the subrequest will not affect the current (parent) request. This option was first introduced in the v0.3.1rc31
release.share_all_vars
specify whether to share all the Nginx variables of the subrequest with the current (parent) request. modifications of the Nginx variables in the subrequest will affect the current (parent) request. Enabling this option may lead to hard-to-debug issues due to bad side-effects and is considered bad and harmful. Only enable this option when you completely know what you are doing.always_forward_body
when set to true, the current (parent) request's request body will always be forwarded to the subrequest being created if the body
option is not specified. The request body read by either ngx.req.read_body() orlua_need_request_body on will be directly forwarded to the subrequest without copying the whole request body data when creating the subrequest (no matter the request body data is buffered in memory buffers or temporary files). By default, this option is false
and when the body
option is not specified, the request body of the current (parent) request is only forwarded when the subrequest takes the PUT
or POST
request method.Issuing a POST subrequest, for example, can be done as follows
res = ngx.location.capture(
'/foo/bar',
{ method = ngx.HTTP_POST, body = 'hello, world' }
)
See HTTP method constants methods other than POST. The method
option is ngx.HTTP_GET
by default.
The args
option can specify extra URI arguments, for instance,
ngx.location.capture('/foo?a=1',
{ args = { b = 3, c = ':' } }
)
is equivalent to
ngx.location.capture('/foo?a=1&b=3&c=%3a')
that is, this method will escape argument keys and values according to URI rules and concatenate them together into a complete query string. The format for the Lua table passed as the args
argument is identical to the format used in the ngx.encode_argsmethod.
The args
option can also take plain query strings:
ngx.location.capture('/foo?a=1',
{ args = 'b=3&c=%3a' } }
)
This is functionally identical to the previous examples.
The share_all_vars
option controls whether to share nginx variables among the current request and its subrequests. If this option is set to true
, then the current request and associated subrequests will share the same Nginx variable scope. Hence, changes to Nginx variables made by a subrequest will affect the current request.
Care should be taken in using this option as variable scope sharing can have unexpected side effects. The args
, vars
, orcopy_all_vars
options are generally preferable instead.
This option is set to false
by default
location /other {
set $dog "$dog world";
echo "$uri dog: $dog";
}
location /lua {
set $dog 'hello';
content_by_lua '
res = ngx.location.capture("/other",
{ share_all_vars = true });
ngx.print(res.body)
ngx.say(ngx.var.uri, ": ", ngx.var.dog)
';
}
Accessing location /lua
gives
/other dog: hello world
/lua: hello world
The copy_all_vars
option provides a copy of the parent request's Nginx variables to subrequests when such subrequests are issued. Changes made to these variables by such subrequests will not affect the parent request or any other subrequests sharing the parent request's variables.
location /other {
set $dog "$dog world";
echo "$uri dog: $dog";
}
location /lua {
set $dog 'hello';
content_by_lua '
res = ngx.location.capture("/other",
{ copy_all_vars = true });
ngx.print(res.body)
ngx.say(ngx.var.uri, ": ", ngx.var.dog)
';
}
Request GET /lua
will give the output
/other dog: hello world
/lua: hello
Note that if both share_all_vars
and copy_all_vars
are set to true, then share_all_vars
takes precedence.
In addition to the two settings above, it is possible to specify values for variables in the subrequest using the vars
option. These variables are set after the sharing or copying of variables has been evaluated, and provides a more efficient method of passing specific values to a subrequest over encoding them as URL arguments and unescaping them in the Nginx config file.
location /other {
content_by_lua '
ngx.say("dog = ", ngx.var.dog)
ngx.say("cat = ", ngx.var.cat)
';
}
location /lua {
set $dog '';
set $cat '';
content_by_lua '
res = ngx.location.capture("/other",
{ vars = { dog = "hello", cat = 32 }});
ngx.print(res.body)
';
}
Accessing /lua
will yield the output
dog = hello
cat = 32
The ctx
option can be used to specify a custom Lua table to serve as the ngx.ctx table for the subrequest.
location /sub {
content_by_lua '
ngx.ctx.foo = "bar";
';
}
location /lua {
content_by_lua '
local ctx = {}
res = ngx.location.capture("/sub", { ctx = ctx })
ngx.say(ctx.foo);
ngx.say(ngx.ctx.foo);
';
}
Then request GET /lua
gives
bar
nil
It is also possible to use this ctx
option to share the same ngx.ctx table between the current (parent) request and the subrequest:
location /sub {
content_by_lua '
ngx.ctx.foo = "bar";
';
}
location /lua {
content_by_lua '
res = ngx.location.capture("/sub", { ctx = ngx.ctx })
ngx.say(ngx.ctx.foo);
';
}
Request GET /lua
yields the output
bar
Note that subrequests issued by ngx.location.capture inherit all the request headers of the current request by default and that this may have unexpected side effects on the subrequest responses. For example, when using the standard ngx_proxy
module to serve subrequests, an "Accept-Encoding: gzip" header in the main request may result in gzipped responses that cannot be handled properly in Lua code. Original request headers should be ignored by setting proxy_pass_request_headers to off
in subrequest locations.
When the body
option is not specified and the always_forward_body
option is false (the default value), the POST
and PUT
subrequests will inherit the request bodies of the parent request (if any).
There is a hard-coded upper limit on the number of concurrent subrequests possible for every main request. In older versions of Nginx, the limit was 50
concurrent subrequests and in more recent versions, Nginx 1.1.x
onwards, this was increased to 200
concurrent subrequests. When this limit is exceeded, the following error message is added to the error.log
file:
[error] 13983#0: *1 subrequests cycle while processing "/uri"
The limit can be manually modified if required by editing the definition of the NGX_HTTP_MAX_SUBREQUESTS
macro in thenginx/src/http/ngx_http_request.h
file in the Nginx source tree.
Please also refer to restrictions on capturing locations configured by subrequest directives of other modules.
Back to TOC
syntax: res1, res2, ... = ngx.location.capture_multi({ {uri, options?}, {uri, options?}, ... })
context: rewrite_by_lua*, access_by_lua*, content_by_lua*
Just like ngx.location.capture, but supports multiple subrequests running in parallel.
This function issues several parallel subrequests specified by the input table and returns their results in the same order. For example,
res1, res2, res3 = ngx.location.capture_multi{
{ "/foo", { args = "a=3&b=4" } },
{ "/bar" },
{ "/baz", { method = ngx.HTTP_POST, body = "hello" } },
}
if res1.status == ngx.HTTP_OK then
...
end
if res2.body == "BLAH" then
...
end
This function will not return until all the subrequests terminate. The total latency is the longest latency of the individual subrequests rather than the sum.
Lua tables can be used for both requests and responses when the number of subrequests to be issued is not known in advance:
-- construct the requests table
local reqs = {}
table.insert(reqs, { "/mysql" })
table.insert(reqs, { "/postgres" })
table.insert(reqs, { "/redis" })
table.insert(reqs, { "/memcached" })
-- issue all the requests at once and wait until they all return
local resps = { ngx.location.capture_multi(reqs) }
-- loop over the responses table
for i, resp in ipairs(resps) do
-- process the response table "resp"
end
The ngx.location.capture function is just a special form of this function. Logically speaking, the ngx.location.capture can be implemented like this
ngx.location.capture =
function (uri, args)
return ngx.location.capture_multi({ {uri, args} })
end
Please also refer to restrictions on capturing locations configured by subrequest directives of other modules.
Back to TOC
context: set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, body_filter_by_lua, log_by_lua*
Read and write the current request's response status. This should be called before sending out the response headers.
ngx.status = ngx.HTTP_CREATED
status = ngx.status
Setting ngx.status
after the response header is sent out has no effect but leaving an error message in your nginx's error log file:
attempt to set ngx.status after sending out response headers
Back to TOC
syntax: ngx.header.HEADER = VALUE
syntax: value = ngx.header.HEADER
context: rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, body_filter_by_lua, log_by_lua*
Set, add to, or clear the current request's HEADER
response header that is to be sent.
Underscores (_
) in the header names will be replaced by hyphens (-
) by default. This transformation can be turned off via thelua_transform_underscores_in_response_headers directive.
The header names are matched case-insensitively.
-- equivalent to ngx.header["Content-Type"] = 'text/plain'
ngx.header.content_type = 'text/plain';
ngx.header["X-My-Header"] = 'blah blah';
Multi-value headers can be set this way:
ngx.header['Set-Cookie'] = {'a=32; path=/', 'b=4; path=/'}
will yield
Set-Cookie: a=32; path=/
Set-Cookie: b=4; path=/
in the response headers.
Only Lua tables are accepted (Only the last element in the table will take effect for standard headers such as Content-Type
that only accept a single value).
ngx.header.content_type = {'a', 'b'}
is equivalent to
ngx.header.content_type = 'b'
Setting a slot to nil
effectively removes it from the response headers:
ngx.header["X-My-Header"] = nil;
The same applies to assigning an empty table:
ngx.header["X-My-Header"] = {};
Setting ngx.header.HEADER
after sending out response headers (either explicitly with ngx.send_headers or implicitly withngx.print and similar) will throw out a Lua exception.
Reading ngx.header.HEADER
will return the value of the response header named HEADER
.
Underscores (_
) in the header names will also be replaced by dashes (-
) and the header names will be matched case-insensitively. If the response header is not present at all, nil
will be returned.
This is particularly useful in the context of header_filter_by_lua and header_filter_by_lua_file, for example,
location /test {
set $footer '';
proxy_pass http://some-backend;
header_filter_by_lua '
if ngx.header["X-My-Header"] == "blah" then
ngx.var.footer = "some value"
end
';
echo_after_body $footer;
}
For multi-value headers, all of the values of header will be collected in order and returned as a Lua table. For example, response headers
Foo: bar
Foo: baz
will result in
{"bar", "baz"}
to be returned when reading ngx.header.Foo
.
Note that ngx.header
is not a normal Lua table and as such, it is not possible to iterate through it using the Lua ipairs
function.
For reading request headers, use the ngx.req.get_headers function instead.
Back to TOC
syntax: headers = ngx.resp.get_headers(max_headers?, raw?)
context: set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, body_filter_by_lua, log_by_lua*, balancer_by_lua*
Returns a Lua table holding all the current response headers for the current request.
local h = ngx.resp.get_headers()
for k, v in pairs(h) do
...
end
This function has the same signature as ngx.req.get_headers except getting response headers instead of request headers.
This API was first introduced in the v0.9.5
release.
Back to TOC
syntax: is_internal = ngx.req.is_internal()
context: set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, body_filter_by_lua*, log_by_lua*
Returns a boolean indicating whether the current request is an "internal request", i.e., a request initiated from inside the current nginx server instead of from the client side.
Subrequests are all internal requests and so are requests after internal redirects.
This API was first introduced in the v0.9.20
release.
Back to TOC
syntax: secs = ngx.req.start_time()
context: set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, body_filter_by_lua*, log_by_lua*
Returns a floating-point number representing the timestamp (including milliseconds as the decimal part) when the current request was created.
The following example emulates the $request_time
variable value (provided by ngx_http_log_module) in pure Lua:
local request_time = ngx.now() - ngx.req.start_time()
This function was first introduced in the v0.7.7
release.
See also ngx.now and ngx.update_time.
Back to TOC
syntax: num = ngx.req.http_version()
context: set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*
Returns the HTTP version number for the current request as a Lua number.
Current possible values are 2.0, 1.0, 1.1, and 0.9. Returns nil
for unrecognized values.
This method was first introduced in the v0.7.17
release.
Back to TOC
syntax: str = ngx.req.raw_header(no_request_line?)
context: set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*
Returns the original raw HTTP protocol header received by the Nginx server.
By default, the request line and trailing CR LF
terminator will also be included. For example,
ngx.print(ngx.req.raw_header())
gives something like this:
GET /t HTTP/1.1
Host: localhost
Connection: close
Foo: bar
You can specify the optional no_request_line
argument as a true
value to exclude the request line from the result. For example,
ngx.print(ngx.req.raw_header(true))
outputs something like this:
Host: localhost
Connection: close
Foo: bar
This method was first introduced in the v0.7.17
release.
Back to TOC
syntax: method_name = ngx.req.get_method()
context: set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, balancer_by_lua*
Retrieves the current request's request method name. Strings like "GET"
and "POST"
are returned instead of numerical method constants.
If the current request is an Nginx subrequest, then the subrequest's method name will be returned.
This method was first introduced in the v0.5.6
release.
See also ngx.req.set_method.
Back to TOC
syntax: ngx.req.set_method(method_id)
context: set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*
Overrides the current request's request method with the method_id
argument. Currently only numerical method constants are supported, like ngx.HTTP_POST
and ngx.HTTP_GET
.
If the current request is an Nginx subrequest, then the subrequest's method will be overridden.
This method was first introduced in the v0.5.6
release.
See also ngx.req.get_method.
Back to TOC
syntax: ngx.req.set_uri(uri, jump?)
context: set_by_lua*, rewrite_by_lua*, access_by_lua*, content_by_lua*, header_filter_by_lua*, body_filter_by_lua*
Rewrite the current request's (parsed) URI by the uri
argument. The uri
argument must be a Lua string and cannot be of zero length, or a Lua exception will be thrown.
The optional boolean jump
argument can trigger location rematch (or location jump) as ngx_http_rewrite_module's rewritedirective, that is, when jump
is true
(default to false
), this function will never return and it will tell Nginx to try re-searching locations with the new URI value at the later post-rewrite
phase and jumping to the new location.
Location jump will not be triggered otherwise, and only the current request's URI will be modified, which is also the default behavior. This function will return but with no returned values when the jump
argument is false
or absent altogether.
For example, the following nginx config snippet
rewrite ^ /foo last;
can be coded in Lua like this:
ngx.req.set_uri("/foo", true)
Similarly, Nginx config
rewrite ^ /foo break;
can be coded in Lua as
ngx.req.set_uri("/foo", false)
or equivalently,
ngx.req.set_uri("/foo")
The jump
argument can only be set to true
in rewrite_by_lua and rewrite_by_lua_file. Use of jump in other contexts is prohibited and will throw out a Lua exception.
A more sophisticated example involving regex substitutions is as follows
location /test {
rewrite_by_lua '
local uri = ngx.re.sub(ngx.var.uri, "^/test/(.*)", "/$1", "o")
ngx.req.set_uri(uri)
';
proxy_pass http://my_backend;
}
which is functionally equivalent to
location /test {
rewrite ^/test/(.*) /$1 break;
proxy_pass