Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
Fetching contributors…

Cannot retrieve contributors at this time

119 lines (87 sloc) 3.558 kb
Description
An nginx module for bringing the power of sh's "echo" command to Nginx's config file.
Installation
Grab the nginx source code from http://nginx.net/ (version 0.7.62
and 0.8.19 have been tested on my side), for example,
the version 0.8.19 and then build the source with this module:
$ wget 'http://sysoev.ru/nginx/nginx-0.8.19.tar.gz'
$ tar -xzvf nginx-0.8.19.tar.gz
$ cd nginx-0.8.19/
$ ./configure --prefix=/opt/nginx \
--add-module=/path/to/echo-nginx-module
Configuration
This is a sample configuration snippet in your nginx.conf file:
http {
...
server {
...
default_type text/plain;
location /hello {
echo 'hello, world';
echo;
echo hello world;
}
location /echoback {
echo "------------------------";
echo_client_request_headers;
}
location /server {
echo "Server name: $server_name";
echo "Server address: $server_addr";
echo "Server port: $server_port";
echo "Server protocol: $server_protocol";
}
location /echo_with_sleep {
echo 'before sleep';
echo_flush; # need this to force flushing outputs before sleep
echo_sleep 1; # sleep for 1 sec here...
echo 'after sleep';
}
...
}
...
}
As you can see, the directives currently provided by this module are
"echo", "echo_client_request_headers", "echo_sleep", and "echo_flush".
And there's also an "echo_blocking_sleep" directive for completeness.
The "echo_sleep" directive is a non-blocking sleep on the server side. So if your
nginx worker process is "sleeping" for one request, it could serve other requests
*at the same time*. On my laptop, with only 1 nginx worker process, my nginx can
achieve 5500+ requests per second at the concurrency level of 10k.
On the other hand, "echo_blocking_sleep" is a blocking version of
"echo_sleep" and it will freeze the whole Nginx worker process. It's
provided merely meant to be compared with "echo_sleep". And you should
never use it for production.
The "echo_flush" directive is needed whenever you need to force flushing the previous
outputs potentially buffered without sending to the client.
After starting the nginx server, one can query it these ways:
$ curl 'http://localhost/hello'
hello, world
hello world
$ curl 'http://localhost/echoback'
------------------------
GET /echoback HTTP/1.1
User-Agent: curl/7.18.2 (i486-pc-linux-gnu) libcurl/7.18.2 OpenSSL/0.9.8g zlib/1.2.3.3 libidn/1.10
Host: localhost
Accept: */*
$ curl 'http://localhost/server'
Server name: localhost
Server address: 127.0.0.1
Server port: 80
Server protocol: HTTP/1.1
$ time curl 'http://localhost/echo_with_sleep'
before sleep
after sleep
real 0m1.018s
user 0m0.012s
sys 0m0.008s
Test suite
This module comes with a Perl-driven test suite. The test cases are
declarative too. Thanks to the Test::Base module in the Perl world.
To run it on your side:
$ cd test
$ PATH=/path/to/your/nginx-with-echo-module:$PATH prove -r t
You need to terminate any nginx processes before running the test
suite if you have changed the nginx server binary.
Contact me
Contact the author by agentzh at gmail dot com whenever you have any questions.
Jump to Line
Something went wrong with that request. Please try again.