config_walkthrough.txt

== link:index.html[Index] -> link:config.html[Configuration]
////
Last checked: 2011/02/17 Cherokee 1.0.21b
////

Configuration Walkthrough
-------------------------

This section briefly describes a typical usage of the administration
web interface provided by
link:other_bundle_cherokee-admin.html[cherokee-admin]. This is the
recommended way of configuring Cherokee. If you are looking for
development information relevant to your automation and scripting
needs, you should refer to the appropriate section, specifically the
link:dev_cherokee.conf.html[cherokee.conf] file specification.

We will first show a quick overview of the available options, followed
by a simple walkthrough. You can learn more about the options in their
specific documentation entries.

You can also take a quick look at the
link:http://www.cherokee-project.com/screencasts.html#intro[introductory
screencast].

image:media/images/screencast.png["introductory screencast",
link="http://www.cherokee-project.com/screencasts.html#intro"]


[[overview]]
Overview
~~~~~~~~
* link:config_index.html[Home]:
  This is the initial screen of cherokee-admin. From here you will be
  able to launch and stop the webserver, change the language of the
  configuration-interface, check a quick view of the server status and
  access relevant support and contact Cherokee-Project links.

* link:config_status.html[Status]:
  This gives access to usage graphs and statistics regarding each
  virtual host and the server as a whole.

* link:config_general.html[General]:
  There are a number of entries that specify the most significant
  configuration options such as the port - or ports - that the server
  will listen to, the default timeout, whether to support keep-alive
  connections, default icon and MIME type definitions and so on.

* link:config_virtual_servers.html[vServers]:
  If you want your web server to work with more than one domain you
  will have to create link:config_virtual_servers.html[Virtual
  servers] other than the `default` one. Each one will have a
  completely independent configuration: paths, behavior, logging
  facilities, etc.

* link:config_info_sources.html[Sources]:
  Define the resources that will be providing information. For
  instance, PHP.

* link:config_advanced.html[Advanced]:
  This is to configure the most complex parameters of the server and how
  it interacts with the operating system. If you are unsure about any of
  the options here, better not mingle with them. Default values should
  work just fine.

[[walkthrough]]
Walkthrough
~~~~~~~~~~~
There is very little set up you must do to have a perfectly functional
web server out of the box. Cherokee's default configuration can be
used to serve any typical static website. For dynamic contents you
should check the wizards, available using the buttons located at the
top section of the left panel in the `vServers` section. This panel
can either display information relevant to every virtual host in your
system, or rule management information specific to a given virtual
host if you are browsing a specific virtual host at the moment. In the
first scenario, the wizards will be able to create new virtual servers
customized for a specific need, while the second case would apply to
the -preexisting- selected virtual host.

If you have a specific need, you should check out the recipes present
in the link:cookbook.html'Cookbook] section of the
documentation. Among other subjects, you can find information about
link:cookbook_optimizations.html[Cherokee optimizations], setting up
efficient link:cookbook_authentication.html[authentication]
mechanisms, or configuring several popular application frameworks.

In this tutorial we will be setting up a system with a couple of
virtual servers, PHP support, some redirection rules and a simple
authentication mechanism.

To follow this walkthrough you need to be running
link:other_bundle_cherokee-admin.html[cherokee-admin]. This is an
administration tool, so you will need system administrator rights. You
can either launch it manually, letting it generate a one-time-password
that you will use to securely connect to the user interface, or you
can follow the recommended way and simply use
link:other_bundle_cherokee-admin-launcher.html[cherokee-admin-launcher].


[[auto-launch]]
Launching the admin through cherokee-admin-launcher
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Assuming you are using a desktop environment on the same host where
Cherokee is installed, you can follow the launcher approach.

----
# cherokee-admin-launcher
Checking TCP port 9090 availability.. OK
Launching /usr/sbin/cherokee-admin.. OK
Connecting.. OK
----

This will automatically open a web browser window pointing at
http://localhost:9090, so you will be able to begin using
`cherokee-admin` right away. If there is no web browser registered in
your system, you will have to forget about using the launcher option,
and run `cherokee-admin` directly. If this is not your case, you can
skip the details for the manual launch.

[[manual-launch]]
Manually launching cherokee-admin
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

----
# cherokee-admin

Login:
  User:              admin
  One-time Password: yp7F1pMhtmD58DbB

Cherokee Web Server 1.0.2 (Jun 19 2010): Listening on port 127.0.0.1:9090,
TLS disabled, IPv6 enabled, using epoll, 4096 fds system limit, max. 2041
connections, caching I/O, single thread
----

Now you can access the administration interface simply by opening your
web browser and visiting
link:http://localhost:9090[http://localhost:9090].
link:other_bundle_cherokee-admin.html[Cherokee-admin] supports several
command line parameters, so you can change the predefined port in case
it was already in use in your system.

The User and One-time Password will be required initially. This is to
prevent other users of the local host from being able to configure the
server unless they have access to the password. This is also very
useful to allow temporary access to a remote administrator.

[[cherokee-admin]]
cherokee-admin
^^^^^^^^^^^^^^

If your user doesn't have enough privileges you will be notified. Also
if no configuration file exists you will be prompted to create one
with the default settings.

image::media/images/admin_noconfig.png[cherokee.conf not present]

The regular template is a good starting point. Right now your web
server will not be running yet. We will only be using two of the
available tabs to adjust some more settings:
link:config_general.html[General] and
link:config_virtual_servers.html[Virtual Servers].

.Sample view
image::media/images/admin_notrunning.png[Server not launched]

Of course, once we're done we will have to apply our changes and
launch the server. The configuration changes must be submitted to the
server, be it automatically if possible or by using the appropriate
form submitting buttons. But no change is reflected in the
configuration file until you actually `Save` the changes. You can be
sure of not damaging anything while playing around with
`cherokee-admin`. At least not until you `Save` the changes by using
the `SAVE` button on the main toolbar. Note that, if the Cherokee
server is running, the applied modifications will only affect the
instance being executed if either `Graceful restart` or `Hard restart`
are selected.  `No restart` simply modifies the configuration file,
but doesn't affect the currently running instance. A `graceful
restart` will preserve the old set-up for any prior connections, while
a `hard restart` will kill every connection and instantly apply the
changes.

.Sample of tab for general settings
image::media/images/admin_general.png[General]

We will only be modifying the *Permissions* tab. It usually is a poor
choice to run services with super user privileges. Set both *User* and
*Group* to `www-data`. Your system might already have another user account
specifically for the purpose of running a webserver. Use that instead
if so, or create the `www-data` user if you don't already have one.

Next is the link:config_virtual_servers.html[vServers] tab.

.Virtual server tab
image::media/images/admin_vserver.png[Virtual server]

When you start, you will only have one virtual server called
`default`. You can begin by cloning it as `example`. This can be
achieved with the `Clone Selected Virtual Server` button, which is
placed in the upper side of the left panel, adjacent to the `Virtual
Servers` label and the `Add New Virtual Server` button.

While you are at it you should also append the following line to your
`/etc/hosts` file.

----
127.0.0.1       example.com example.net example.org
----

This is to allow your system to resolve requests for `example.net` or
`example.org` locally. Of course, once you are out in the Wild (you
know, in the Internet), you will need proper DNS records.


*`default`*::
We will leave default untouched except for the document root path,
which we will change. This is done through the `Basics` tab.
You could create by hand '/var/www/default' for example and set it up
in the appropriate field.
+
If you access your machine by IP or a DNS resolution points in that
direction for any particular domain, the contents of your document
root directory will be exposed ('/var/www' by default unless you
changed that during the build process). Whenever there is no match for
a virtual server specifically defined in your list, the `default`
virtual server will be the one responding.
+
To enable PHP support you will have to select the desired virtual
server, access the `Behavior` tab, and click on the `Rule Management`
button. This will replace the `Virtual Servers` panel with a `Rules`
panel that applies only to the selected virtual server. The `Add`
button now will give you access to manually adding rules or to the
list of wizards. You will find PHP under the `Languages`
category. Simply point and click, and this will make the required
adjustments. Any file with the 'php' extension will be served after being
processed by the PHP interpreter. You should have `php-cgi` installed
in your system for this, though. Don't worry too much about it: the
wizard will give you instructions and will notify any further
requirements not yet satisfied by your system.
+
Any requested file would be sent. If a directory is requested, a file
called index.php or index.html will be served if it is present (the
search will be performed in that order; you can configure this in the
`Basics` tab). If not, a directory listing is offered instead. If you
would want to prevent this behavior but would like to keep up
serving whatever content is requested directly, simply change the
`List & Send` handler for the `Static Content` handler.


*`example`*::
Make sure to select the correct virtual server in the `Virtual
Servers` panel displayed in the left side of the interface. If the
panel was titled `Behavior` instead, make sure it applies to the
desired virtual server. This can be easily checked, since it is
displayed as a title in the upper part of the contents area, which is
everything else on the screen that doesn't belong to the panel. If you
are inside another virtual server, just click on the host name to
restore the `Virtual Servers` panel. If you have already selected the
correct virtual host, access the `Behavior` tab and click on the `Rule
Management` button to see the `Behavior` panel that will enable you to
modify the rules of your virtual host.
+
For now this virtual server is an exact copy of the untouched
`default` virtual server. Create some new directories by hand:
'/var/www/example' and '/var/www/example/auth', and configure the
first one as default document root path. Just access the `Basics` tab,
and set the `Document Root` field to '/var/www/example'.
+
Next, setup `example.net` and `example.com` in the `Host match` tab,
so that requests to both domains can be managed byt your `example`
virtual server. To do so just choose `Wildcards` as method, and add
both names as wildcard strings.  We'll erase everything in the
*Targets* list within the `Behavior` tab, except the `default`
rule. With this we'll only be able to serve static content.
+
Accessing the URL http://example.org should now show a list of
available files and directories under '/var/www/example/'.
+
Now lets password protect the `auth` directory. Add a new
`Directory`-type rule pointing to `/auth`. Then, through the
`Security` tab, add a `Validation Mechanism` under `Authentication`.
+
The field `Realm` is mandatory. Lets set it as `Secured Area`.
If you choose `PAM` and impose no more restrictions, only users with
a local account in your system will be able to access the secured zone
at http://example.net/auth
Other mechanisms would work in a similar fashion but with their
specific requirements. For example, had you chosen `Plain text file`
instead of PAM you would have had to specify a `Password File`.
For example, it could have been `/var/www/passwords.txt` with the
following contents:
+
----
testuser1:password1
testuser2:password2
----
+
Note that the rule must not be flagged as `Final`, or else no other
rules will be applied afterwards and no access will be given to any
contents. This is because at this point we have not defined any
handler for the `/auth` rule and nothing would be served.
+
Refer to the link:cookbook_authentication.html[Cookbook] for detailed
examples on the different options.
+
Lastly, lets configure a redirection rule by choosing a "Regular
Expression" as the new rule type. Type the following regex: `^/rss.*$`
Then, configure the `Handler` tab with the following settings.

. *Handler*: `Redirection`
. *Type*: `External`
. *Regular Expression*: you can leave it empty or specify a series of
  new matches. The shot has the base regex for ilustration purposes:
  `^/rss.*$`
. *Substitution*: this will be the target: `/feed`

+
.Redirection handler
image::media/images/admin_rule_regex.png[Regex]
+
And voilà! A request of the form http://example.net/rss2 would be
redirected to http://example.net/feed
+
Any rule type can be used with the redirection handler: Directory,
Extensions, Regular Expression, etc. In this case the fact that the
rule type is "Regular Expression" affects only slightly. Refer to the
appropriate section of the documentation,
link:modules_handlers_redir.html[Redirection Handler], for more
details.
+
Note that the `Type` of the *Redirection* is `External`. This means
the server will instruct the requesting web client to fetch the
redirected URL, which in turn means the client will always know what
the final URL is. This also means the redirection can be done to
servers other than your own. If it were internal, the redirection
would be invisible (not showing the target URL), but it would be
limited to the same `virtual server`.
+
A much more general redirection could be one using these values:
+
*Regular Expression*;;
  `^/(.*)$`
+
*Substitution*;;
  http://www.example.com/example.net/$1

+
Note that this rule would have to be external since `example.com` is
not among the domains managed by our configuration.
+
This would redirect every petition to a site hosted under
`http://example.com/example.net`. For instance, the request for
`http://example.net/image.jpg` would return
`http://example.com/example.net/image.jpg`.
+
If you need more details, Check out the documentation for the
link:modules_handlers_redir.html[redirection] handler.