Configuration Options

Home | Getting Started | API | Elements | Actions | Validators | Handlers | Configuration Options | Plugins | Mobile | Troubleshooting | About

Configuration Options

Table of Contents

1 Configuration Overview

Nitrogen configuration is done in a handful of different configuration files, and also varies slightly between backend servers. All of these files are found in the etc/ directory of your Nitrogen installation. Below we will go though the configuration options provided within each.

2 Univeral Configuration Files (configs available in all Nitrogen intallations)

etc/vm.args

This is the file that sends sends options to the Erlang Virtual Machine. Basically, any of the usual erl startup parameters will work here. You'll define some standard options here, such as the node name, (-name), erlang cookie (-cookie), some miscellaneous environment options, and of course, to tell the system to launch the Nitrogen application upon startup.

Here is the complete text of the default vm.args:

## Name of the Nitrogen node
-name nitrogen@127.0.0.1

## Cookie for distributed erlang
-setcookie nitrogen

## Heartbeat management; auto-restarts VM if it dies or becomes unresponsive
## (Disabled by default..use with caution!)
##-heart

## Enable kernel poll and a few async threads
+K true
+A 5

## Increase number of concurrent ports/sockets
-env ERL_MAX_PORTS 4096

## Tweak GC to run more often
-env ERL_FULLSWEEP_AFTER 10

## Include .beam files for site.
-pa ./site/ebin

## Include libraries and rebar dependency paths
-env ERL_LIBS ./lib

## Run code at startup.
-eval "application:start(nitrogen)"

etc/app.config

This file contains the configuration options for the Nitrogen application itself. It uses the standard Erlang config syntax: (a proplist of {applicationName, AppOptions}).

For the nitrogen application, the following AppOptions are recognized:

signkey - (String)
signkey is the term that Nitrogen usees to "sign" the state information that's sent to the client, and then resent back to the server with requests. While Nitrogen's state info is not encrypted, this is used to help verify that the term generated is actually generated by Nitrogen, to prevent client-side tampering with the state.
session_timeout - (integer)
session_timeout is the number of minutes a session can be idle before Nitrogen expunges it from memory
cookie_name - (string)
cookie_name is the name of the HTTP Cookie to be used as the sesson tracking cookie. This is not to be confused with the Erlang Cookie, which is defined in the vm.args file above.

Also typically defined in the app.config file here is the SASL Configuration, the default of which for Nitrogen can be seen here:

module_prefix - (string)
This provides the user with the ability to route all Nitrogen web requests to modules that all share the same prefix. For example, if module_prefix was set to "abc", then a request for /mypage would be directed to the module "abc_mypage".
module_aliases - (proplist)
This provides the template system with a default list of module aliases to use within Nitrogen templates. (See the Template Element for more details).
{sasl, [
    {sasl_error_logger, {file, "log/sasl-error.log"}},
    {errlog_type, error},
    {error_logger_mf_dir, "log/sasl"},      % Log directory
    {error_logger_mf_maxbytes, 10485760},   % 10 MB max file size
    {error_logger_mf_maxfiles, 5}           % 5 files max
]}

plugins.config

The Nitrogen plugin system is capable of automatically including plugin header files, along with importing static web resources (javascript files, images, etc). This can be done, and can be configured to work with just about any Nitrogen configuration, including non-standard configurations, such as manually importing Nitrogen as a dependency.

The plugins.config file contains information for how Nitrogen will import plugins.

It's a standard Erlang config file that takes three arguments.

plugins_hrl - (path string)
Tell the Nitrogen plugin importer where to put the generated plugins.hrl file for the purposes of including plugin elements into your application. (default: ="./site/include/plugins.hrl"=)
static_dir - (path string)
Tell the plugin system where you wish to put your plugins' static resources. (default: ="./site/static/plugins"=)
copy_mode - (copy|link)
Tell the plugins system how to include any static resources. copy will copy the entire contents of your plugins' static directories, while link will merely create a symlink. If you work primarily with Linux or OSX, you can probably get away with using link, while if you use Windows, you should stick with copy. (default: copy).

Here's the complete text of the default plugins.config:

%% vim: ts=2 sw=2 et ft=erlang
%% Nitrogen Plugin Installer Configuration
%%
%% This will tell the plugin-installer script where to put Nitrogen plugin
%% information it finds.
%%

%% plugins_hrl tells the Nitrogen plugin installer where to put the .hrl file
%% containing links to all your plugins' respective header files.
%% The default is "./site/include/plugins.hrl".
{plugins_hrl, "./site/include/plugins.hrl"}.

%% static_dir tells the Nitrogen plugin installer where to copy your plugins'
%% static resources (images, js, etc).
%% The default is "./site/static/plugins"
{static_dir, "./site/static/plugins"}.

%% copy_mode determines if static resources are copied or merely symlinked.
%% Keep in mind, symlinks do not work on Windows, so "copy" is the default.
%% Valid values are the atoms 'copy' or 'link'
%% Default: copy
{copy_mode, copy}.

More about Plugins

The plugin system has its own complete documentation along with a sample plugin stub for creating your own plugins.

We strongly advise reading the Plugin Documentation.

Notable mention: rebar.config

The standard for Erlang distribution and building is the use of Basho's rebar. Nitrogen takes advantage of this for simplifying the process of making releases and compiling Nitrogen even after a release is built and deployed.

Generally, the main reason one would want to customize their installation is by adding additional dependency packages. For example, if you wanted to include the erlware_commons package for improved date parsing and formatting, or the erls3 package to give your app an interface to Amazon S3, you would typically do it by adding the dependencies to rebar.config then running make in your Nitrogen installation.

By default, the only dependencies are the core dependencies for Nitrogen: nitrogen_core, nprocreg, sync, simple_bridge, and a webserver (Yaws, Cowboy, etc).

Also contained within the rebar.config are a handful of other compilation options: minimum Erlang version, where dependencies go, and debugging options.

Below is the rebar.config file when used with webmachine:

{sub_dirs, [
    "site",
    "deps"
]}.

{require_otp_vsn, "R13B04|R14|R15"}.

{cover_enabled, true}.

{erl_opts, [debug_info, fail_on_warning]}.

{deps_dir, ["lib"]}.

{deps, [
    {webmachine, "1.8.*", {git, "git://github.com/basho/webmachine.git", {tag, "webmachine-1.8.1"}}},

    {nitrogen_core, "2.1.*", {git, "git://github.com/nitrogen/nitrogen_core", "HEAD"}},
    {nprocreg,      "0.2.*", {git, "git://github.com/nitrogen/nprocreg", "HEAD"}},
    {simple_bridge, "1.2.*", {git, "git://github.com/nitrogen/simple_bridge", "HEAD"}},
    {sync,          "0.1.*", {git, "git://github.com/rustyio/sync.git", "HEAD"}}
]}.

To add the above mentioned dependencies (erlware_commons and erls3), edit the rebar.config file and modify the deps list to look like this:

{deps, [
    {webmachine, "1.8.*", {git, "git://github.com/basho/webmachine.git", {tag, "webmachine-1.8.1"}}},

    %% Add our two new dependencies below
    {erls3, "1.9.*", {git, "git://github.com/shane42/erls3.git", "HEAD"}},
    {erlware_commons, ".*", {git, "git://github.com/erlware/erlware_commons.git", "HEAD"}},

    {nitrogen_core, "2.1.*", {git, "git://github.com/nitrogen/nitrogen_core", "HEAD"}},
    {nprocreg,      "0.2.*", {git, "git://github.com/nitrogen/nprocreg", "HEAD"}},
    {simple_bridge, "1.2.*", {git, "git://github.com/nitrogen/simple_bridge", "HEAD"}},
    {sync,          "0.1.*", {git, "git://github.com/rustyio/sync.git", "HEAD"}}
]}.

Then run make from the root of your Nitrogen installation. This will download the new dependencies and install them into the lib directory of your installation.

3 Server-specific Configuration (in alphabetical order)

Common Configurations:

In most of the configurations below, you'll find some general running themes. Almost all of the servers take same basic set of core instructions to for initialization and address binding, as well as some way to notify Nitrogen and SimpleBridge where to find static files. These universal configuration settings are below:

bind_address - (String)
The string of the IP address to bind. If set to "0.0.0.0" or left blank, it'll bind to all available addresses. (Default: ="0.0.0.0"= )
port - (Number)
The port number to bind. (Default: 8000)

About Ports and Linux: While port 80 is the standard HTTP port,port 80 is a privileged port in a Linux/Unix environment. This means that in order for Erlang to bind to port 80, it will need to be run with root privileges. This is generally unadvised. Instead, we recommend using a lightweight reverse proxy (such as nginx) in front of Nitrogen. Doing so will allow you to run Nitrogen with standard user privileges (for better system security), while presenting your Nitrogen website on the expected port 80.

On some variants of Linux, it is possible to bind Nitrogen to port 80 without running as root. This is accomplished with the use of the setcap application (which may need to be installed from your distro's package system).

An example of setcap being run on your Erlang app:

sudo setcap cap_net_bind_service+ep ./erts-5.9.2/bin/beam
sudo setcap cap_net_bind_service+ep ./erts-5.9.2/bin/beam.smp

This will give the beam and beam.smp programs privileges to bind to privileged ports (ports under 1024).

server_name - (Erlang term)
What to name the server. (Default: nitrogen)
document_root - (String)
The root of the location of static resources (ie, stylesheets, javascript files, images, etc). This will be passed to simple_bridge for the serving of static files. (Default: ="./site/static"= )

Note: this is relative to the root of the Nitrogen installation.

Cowboy: etc/cowboy.config

Cowboy is the web server made by Loïc Hoguin. The configuration for Cowboy provided with Nitrogen isn't the official configuration file, but it works for our purposes. The options we provide are read by Nitrogen and passed to the Cowboy server upon initialization.

The default Cowboy configuration is as follows:

[
    {cowboy,[
        {bind_address,"0.0.0.0"},
        {port,8000},
        {server_name,nitrogen},
        {document_root,"./site/static"},
        {static_paths, ["js/","images/","css/","nitrogen/"]}
    ]}
].
  • static_paths - (List of Strings)

    This setting will be used to determine if a requested resource should be handled by Nitrogen and simple_bridge, or if it should just be immediately served directly by the Cowboy server. (Default: ["js/","images/","css/","nitrogen/"])

    Note 1: This is relative to the document_root above. So requests for js/ will be served from ./site/static/js/ (using the default above).

    Note 2: it is strongly recommended to catch static files with the static_paths setting. simple_bridge does not serve large static files in an optimal way (it loads the files into memory completely before sending).

Inets: etc/inets.config and etc/inets_httd.erlenv

Inets is the web client and server included with the standard Erlang Install, and we use the Inets Web server as the "simple" solution for Nitrogen. Inets isn't as feature-rich as the other popular Erlang webservers, and because of this, we only recommend using Inets for development purposes, since it doesn't require any additional installation.

Further, the Inets configuration is broken into two different files, one for the inets application itself, and one for the httpd server included in Inets.

etc/inets.config

This is the file for configuring the inets application itself. By default, we simply use this file to tell the application to start the httpd and load the specified configuration file.

The default inets.config provided with Nitrogen is as follows:

[{inets, [
    {services, [
        {httpd, [
            {proplist_file, "./etc/inets_httpd.erlenv"}
        ]}
    ]}
]}].

Note that basically all it does it tell inets to load the inetshttpd.erlenv proplist file, using the proplist_file option.

etc/inets_httpd.erlenv

This file does the heavy lifting of configuring our inets configuration.

[
    {port, 8000},
    {bind_address, {0,0,0,0}},
    {server_name, "nitrogen"},
    {server_root, "."},
    {document_root, "./site/static"},
    {error_log, "./log/inets.log"},
    {modules, [nitrogen_inets]},

    {mime_types, [
        {"css", "text/css"},
        {"js", "text/javascript"},
        {"html", "text/html"}
    ]}
].
bind_address - (IP Address as a 4-tuple)
Note that the bind_address for Inets is different than for the rest of the servers in that it expects the address to be in the form of a 4-tuple for example, instead of specifying the string (ie ="12.34.56.78"=, you would specify {12,34,56,67}).
error_log - (String)
The name of the file to store the inets logs.
modules - (List of module names)
For each request, Erlang will attempt to call ModuleName:do/1 for each specified module. Typically, we just put in the atom nitrogen_inets as that's the default Nitrogen entry point for inets.
mime_types - ([{Extension,Mimetype},…])
This is simply a list of the Mime Types you wish to support along with the extensions that trigger those mime types. By default, it supports css, javascript, and html files. More types will have to be added by the user.

Mochiweb: etc/mochiweb.config

Mochiweb is a webserver written by Bob Ippolito. It's a very lightweight webserver and very easy to configure.

The default configuration file for Mochiweb provided by Nitrogen is as follows:

[{mochiweb, [
    {bind_address, "0.0.0.0"},
    {port, 8000},
    {server_name, nitrogen},
    {document_root, "./site/static"},

    %% Max Request size of 25MB. While this is a mochiweb env_var,
    %% it's actually only used in simple_bridge
    {max_request_size, 26214400}
]}].
max_request_size - (Integer)
Tells Mochiweb (in particular, it tells SimpleBridge) what the maximum request size to be honored. This is in bytes. The current default is 25 MB maximum request size.

Webmachine: etc/webmachine.config

Webmachine is a web server written by Basho (the makers of Riak), and it provides functions to specify detailed dispatch rules.

The basic config file provided for Webmachine is very simple and minimal (it's basically the same as the one for cowboy)

[{webmachine, [
    {bind_address, "0.0.0.0"},
    {port, 8000},
    {document_root, "./site/static"},
    {server_name,nitrogen},
    {static_paths, ["js/","images/","css/","nitrogen/"]}

]}].
static_paths - (List of Strings)
Used to determine if a requested resource should be handled by Nitrogen, or if it should just be immediately served directly by Webmachine. (Default: ["js/","images/","css/","nitrogen/"])

Note 1: This is relative to the document_root above. So requests for js/ will be served from ./site/static/js/ (using the default above).

Note 2: it is strongly recommended to catch static files with the static_paths setting. simple_bridge does not serve large static files in an optimal way (it loads the files into memory completely before sending).

More Webmachine Dispatch Rules: site/src/nitrogen_sup.erl

Webmachine also provides a dispatch table to allow you to specify how requests are handled (beyond the basics covered by Nitrogen and the configuration above). If you're interested in diving into that, check out the site/src/nitrogen_sup.erl file in your Nitrogen installation.

Yaws: etc/yaws.config

Yaws is a high performance webserver created by Claes Wikstrom and is a unique addition to the Nitrogen's supported webserver line-up because it's one of the few that uses Apache-style configuration instead of the more usual Erlang proplist config files.

etc/yaws.config

This file is just tells Yaws where to load the actual configuration file, which you can probably deduce.

[{yaws, [
    {conf, "./etc/yaws.conf"}
]}].
etc/yaws.conf
logdir = ./log
<server mydomain.org>
    port = 8000
    listen = 127.0.0.1

    #the static code to be served directly by yaws is found in ./site/static
    docroot = ./site/static

    # tell yaws to pass control to the nitrogen_yaws module
    # (specifically nitrogen_yaws:out/1) for all requests except for any request
    # that starts with "images/", "nitrogen/", "css/", or "/js".
    # Bear in mind, however, the caveat to this performance improvement:
    # this means that you cannot have any pages called "nitrogen_xxx" or "css_yyy" because
    # the yaws config will see the "exclude_paths" rule below and completely ignore nitrogen.
    # Should you wish to have yaws handle any more static files, for example, if you added
    # a videos directory in site/static/, you can simply add "videos" to the end of the list
    # Ex: appmods = </, nitrogen_yaws exclude_paths images nitrogen css js videos>
    appmods = </, nitrogen_yaws exclude_paths images nitrogen css js>
</server>

You can find the complete documentation for the yaws.conf file on the official website, but for the sake of convenience, here's the a brief description of the default one provided by Nitrogen.

logdir
tells where to store the Yaws log files
<server mydomain.org> [...] </server>
Defines a virtual server. For use with Nitrogen, we recommend only specifying one. mydomain.org in our example is simply the name of the virtual server, and is not used for anything beyond a naming scheme.
port
The port to listen on.
listen
Which IP address to listen on.
docroot
The location of the static files relative to the Nitrogen installation
appmods = </, nitrogen_yaws exclude_paths images nitrogen css js>
While quite long and dense with information, this configuration setting tells Yaws to send all requests to the Erlang module nitrogen_yaws, except for any requests that start with /images, /nitrogen, /css, or /js, which will instead be handled by Yaws directly.

4 Additional Configuration

nginx - A lightweight reverse proxy

Nginx is high performance, lightweight web server and reverse proxy that is commonly used for load balancing, rewrite rules, SSL certificates, and more.

Here's a sample configuration (this assumes a standard Ubuntu configuration):

/etc/nginx/nginx.conf
user www-data;
worker_processes  1;

error_log  /var/log/nginx/error.log;
pid        /var/run/nginx.pid;

events {
            worker_connections  4096;
}

http {
        include       /etc/nginx/mime.types;
        default_type  application/octet-stream;

        access_log      /var/log/nginx/access.log;

        sendfile        on;

        keepalive_timeout  10;
        tcp_nodelay        on;

        gzip  on;

        proxy_set_header X-Forwarded-Host $host;
        proxy_set_header X-Forwarded-Server $host;
        proxy_set_header Host $host;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;


        include /etc/nginx/conf.d/*.conf;
        include /etc/nginx/sites-enabled/*;
}
Non-SSL Sample: /etc/nginx/sites-enabled/my_site
server {
        listen   80;
        server_name  mysite.com www.mysite.com;
        access_log  /var/log/nginx/mysite.com.access.log;
        location / {
                proxy_pass http://127.0.0.1:8000;
        }
}
SSL-Only Sample: /etc/nginx/sites/enabled/my_secure_site

This configuration will server only SSL. It will redirect all requests from the HTTP port (port 80) to the HTTPS port (port 443) and load the certificates

# My config for a site that I only want serving SSL content.
server {
        listen   80;

        server_name www.mysite.com, mysite.com;
        access_log  /var/log/nginx/mysite.com.access.log;

  # rewrite all requests to be SSL
        rewrite ^(.*) https://$host$1 permanent;
}

server {
        listen 443;
        server_name mysite.com www.mysite.com
        access_log /var/log/nginx/mysite.ssl.access.log;

        ssl on;

        ssl_certificate ssl/mysite/mysite.com.crt;
        ssl_certificate_key ssl/mysite/mysite.com.key;
        ssl_client_certificate ssl/mysite/ca.crt;

        location / {
        # This installation is running on port 8021, as you can plainly see.
          proxy_pass http://127.0.0.1:8000;
        }
}

Date: 2013-10-18 07:27:00 CDT

Author: Jesse Gumm (@jessegumm)

Org version 7.8.02 with Emacs version 23

Validate XHTML 1.0

Comments

Note:To specify code blocks, just use the generic code block syntax:
<pre><code>your code here</code></pre>


comments powered by Disqus