A Resin server can serve many virtual hosts, each with
its own servlets and documents. The configuration is flexible,
allowing dynamic host deployment in a hosts directory
or using explicit <host> tags for additional control and security,
and compatibility with existing Apache sites, enabling easy upgrades
and evaluation for PHP servers to Quercus.
Virtual hosts are multiple internet domains served by the same
Resin server. Because one JVM handles all the domains, its more memory
and processing efficient, as well as sharing IP addresses. With Resin,
adding virtual hosts can as easy as creating a directory
like /var/www/hosts/foo.com and setting up the DNS name.
Explicit virtual host is also possible to match existing layouts, like
matching a /var/www/htdocs configuration when migrating
a PHP mediawiki or wordpress site to use Quercus
for security and performance.
The virtual host will contain one or
more web-apps to serve the host's contents.
Simple sites will use a fixed root webapp, like the Apache-style
/var/www/htdocs. More complicated sites can use a
webapps-style directory.
Each virtual host belongs to a Resin <cluster>, even if the cluster has only a single server.
For example, a Resin server might manage both the
www.gryffindor.com and www.slytherin.com domains,
storing the content in separate directories (/var/www/gryffindor and
/var/www/slytherin), and using a single IP address for both domains.
In this scenario, both www.gryffindor.com and www.slytherin.com are
registered with the standard domain name service registry as having the IP
address 192.168.0.13. When a user types in the url
http://www.gryffindor.com/hello.jsp in their browser,
the browser will send the HTTP request to the IP address
192.168.0.13 and send an additional HTTP header for the
gryffindor host, "Host: www.gryffindor.com". When Resin receives the request
it will grab the host header, and dispatch the request to the configured
virtual host.
Resin can deploy virtual hosts automatically by scanning a host
deployment directory for virtual host content. Each sub-directory in
the hosts directory will cause Resin to create a new
virtual host. To customize the configuration, you can
add a host.xml in the host's root directory for shared
databases, beans or security, or to add
<host-alias> names.
You can add hosts dynamically to a running server just by creating
a new host directory. Resin periodically scans the hosts
directory looking for directory changes. When it detects a new host directory,
it will automatically start serving files from the new virtual hosts.
If you add a default directory in the hosts,
Resin will use it to serve all unknown virtual hosts. The default host is
handy for simple servers with only a single virtual host and for sites
where the virtual host is handled in software, like Drupal. If the
default directory is missing, Resin will
return 404 Not Found for any unknown virtual hosts.
Often, the same virtual host will respond to multiple names, like
www.slytherin.com and slytherin.com. One
name is the primary name and the others are aliases. In Resin, the
primary name is configured by the <host-name> tag and aliases are configured
by <host-alias>. In a dynamic
host configuration, the directory name is used as the host-name
by default, and aliases are declared in the host.xml.
Since the host.xml is shared for all web-applications in
the host, you can also use it to configure shared resources like security
logins, shared databases, and shared resources.
The <host-deploy> tag configures
the dynamic virtual hosting specifying the directory where Resin should
scan for virtual hosts. Because Resin does not automatically add default
configuration, you will need to also add configuration for the
host.xml, app-default.xml and
web-app-deploy. Although it's a bit more verbose, the
no-default rule makes Resin more secure and debuggable. If an item like
a <web-app> is missing, Resin will return 404 Not Found for
security. Because all configuration is explicit, it's ultimately traceable
to the resin.xml which makes debugging more reliable.
Shared host configuration goes in the
<host-default> tag. In this
case, we've added an optional host.xml for configuration,
an access log in log/access.log and a standard
webapps directory. The standard servlets and file handling
come from the app-default.xml file. If you omit either the
app-default.xml or the webapps, you will see 404 Not Found for
any requests.
The example below is a complete, working resin.xml listening
to HTTP at port 8080. The cluster
consists of a single server. It includes a <> to help debugging
the configuration. Many sites will omit the error-page to
hide configuration details in case an error occurs on a live site.
Any directory created in ${resin.root}/hosts will now
become a virtual
host. You can also place a .jar file in ${resin.root}/hosts, it is expanded to
become a virtual host.
Jar libraries and class files that are shared amongst all webapps in the host
can be placed in lib and classes subdirectories of the host:
More information is available in the configuration documentation for <host-deploy> and <host-default>.
In a more structured site, you can take complete control of the
virtual host configuration and configure each virtual host explicitly.
Existing sites wanting to upgrade to Resin or sites with extra security needs
may prefer to configure each <host> in the resin.xml. For
example, a PHP Drupal site evaluating Quercus to
improve performance and security might use the explicit <host> to
point to the existing /var/www/htdocs directory.
In the explicit configuration, each virtual host has its own host block. At the very least, each host will define the id specifying the host name and a root web-app. A is often used to provide a host specific root for logfiles.
As with the dynamic hosting, servlets and web-apps must be configured
either in a <host-default> or explicitly. If they are missing, Resin
will return a 404 Not Found for security. The host id=""
is the default host and will serve any request that doesn't match other
hosts. If you don't have a default host, Resin will return a
404 Not Found for any unknown host.
The following sample configuration defines an explicit virtual hosts
www.slytherin.com and a default host, each with its own
root directory, access-log and a single explicit
<web-app> in the htdocs
directory. The default virtual host is configured just like a typical
Apache configuration, so it can be used to upgrade an Apache/PHP site
to use Quercus for security and performance.
Browsing http://gryffindor.caucho.com/test.php will look for /var/www/htdocs/test.php.
Browsing http://slytherin.caucho.com/test.php will look for /var/slytherin/htdocs/test.php.
In some ISP setups, it may make sense to assign a server for each virtual host. The isolation of web-apps may not be sufficient; each host needs a separate JVM. In this configuration, each <host> belongs to its own <cluster> and has a dedicated <server>. Normally, this configuration will operate using load-balancing, so the load-balance server will dispatch requests as appropriate.
For further security restrictions, see the watchdog section. ISPs can also use the watchdog to assign different <user-name> values for each host and can even create chroot directories for each JVM.
A front-end web server receives all requests, and is configured to dispatch to back-end Resin server that correspond to the host name.
Each host is placed in its own <cluster> with a dedicated <server>. Since the server listens to a TCP port for load-balancing and clustering messages, each server on the maching needs a different server port.
In this example, the virtual hosts www.gryffindor.com and
www.slytherin.com each get their own server. The backend
clusters have their own virtual host. The frontend load-balancer dispatches
the <resin:LoadBalance> tags
to the backend.
This example is split into two blocks to emphasize the frontend and backend. Typically, they will both actually be in the same resin.xml to ensure consistency.
Each back-end server is started separately:
The host-specific back-end servers are ready to receive requests on their server ports. A third Resin server can be used as the front-end load-balancer. It receives all requests and dispatches to the back-end servers.
The Resin web server is configured using rewrite with a <resin:LoadBalance directive to dispatch to the back-end server. A cluster is defined for each back-end host, so that the <load-balance> knows how to find them.
The front-end server JVM is started similar to the back-end JVMs:
With Windows, each JVM is installed as a service. Service is installed using setup.exe graphical utility. It's possible to install a number of Resin services each using a unique name. The name will need to be supplied into 'Service Name' field.
You will either need to reboot the machine or start the service from the Control Panel/Services panel to start the server. On a machine reboot, NT will automatically start the service.
There is a bug in many JDKs which cause the JDK to exit when the administrator logs out. JDK 1.4 and later can avoid that bug if the JDK is started with -Xrs.
The virtual host name can be configured by
an explicit <host-name>,
a <host-alias>,
a <host-alias-regexp>,
by the <host> tag or
implicitly by the <host-deploy>.
For explicit configuration styles, the host name and alias
configuration will generally be in the resin.xml. For dynamic configuration,
the host aliases will typically be in an included host.xml inside
the host directory.
The default host catches all unmatches hosts. Simpler sites will
use the default host for all requests, while security-conscious sites
may remove the default host entirely. If the default host is not configured,
Resin will return a 404 Not Found.
The host.xml is an optional file where virtual hosts can
put host-common configuration. The host.xml
is a good place for shared resources like authentication, database pools
or host-wide beans and services. It's also a
location for the <host-alias> in a dynamic hosting configuration.
The host.xml is configured in a <host-deploy>
or <host-default> by adding a
<resin:import> tag
specifying the host.xml name and location. Because the <host-default>
applies the <resin:import> to every virtual host, it becomes a common
system-wide configuration file.
Hosts must define web-apps in order to
serve files, servlets, or PHP pages. If the host is missing all
webapps, Resin will return 404 Not Found for all requests
make to the host.
Both explicit <web-app> and dynamic web-app-deploy tags are used to configure webapps. The explicit style is generally used for Apache-style configuration, while the dynamic style is generally used for Java app-server .war configuration.
Remember, Resin's default servlets like the file, JSP, and PHP servlets
also need to be defined before they're used. So all Resin configuration
files need to have a <resin:import> of the conf/app-default.xml
configuration file either in the <cluster> or in a
shared <cluster-default>. If the app-default.xml is missing, Resin will
not serve static files, JSP, or PHP, and will not even look in the
WEB-INF for resin-web.xml, classes, or lib.
While Resin's virtual hosting is primarily aimed at named-based virtual hosts, it's possible to run Resin with IP-Based virtual hosts.
With IP virtual hosting, each <http> block is configured with the virtual host name. This configuration will override any virtual host supplied by the browser.
Resin's virtual hosting understands host names encoded using rfc3490 (Internationalizing Domain Names in Applications). This support should be transparent. Just specify the virtual host as usual, and Resin will translate the brower's encoded host name the unicode string.
Support, of course, depends on the browser. Mozilla 1.4 supports the encoding.
A common configuration uses virtual hosts with Apache or IIS. As usual, Apache or IIS will pass matching requests to Resin.
The Resin JVM configuration with Apache is identical to the standalone configuration. That similarity makes it easy to debug the Apache configuration by retreating to Resin standalone if needed.
The ServerName directive in Apache with the UseCanonicalName can be used to select a canonical name for the virtual host virtual hosting work. When Apache passes the request to Resin, it tells Resin the ServerName. Without the ServerName, Apache will use the "Host:" header in the HTTP request to select which host to serve.
The host-specific back-end JVMs are ready to receive requests on their server ports. Apache is the front-end server, and is configured to dispatch to the appropriate back-end Resin JVM for the host:
When you restart the Apache web server, you can look at http://gryffindor/caucho-status and http://slytherin/caucho-status to check your configuration. Check that each virtual host is using the server address and port that you expect.
During development and testing, it is often inconvenient or impossible to use real virtual host names that are registered as internet sites, and resolve to an internet-available IP address. OS-level features on the test client machine can be used to map a virtual host name to an IP address.
For example, developers often run the Resin server and the test client (usually a browser) on the same machine. The OS is configured to map the "www.gryffindor.com" and "www.slytherin.com" names to "127.0.0.1", pointing these host names back to computer that the client is running on.
Unix users edit the file /etc/hosts:
Windows user edit the file C:\WINDOWS\SYSTEM32\DRIVERS\ETC\HOSTS:
Specifies ear expansion.
ear-deploy can be used in web-apps to define a subdirectory for ear expansion.
<error-page> defines a web page to be displayed when an error occurs outside of a web-app. Note, this is not a default error-page, i.e. if an error occurs inside of a <web-app>, the error-page for that web-app will be used instead.
See webapp: error-page.
<host> configures a virtual host. Virtual hosts must be configured explicitly.
It is recommended that any <host> using a regexp include a <host-name> to set the canonical name for the host.
<host-alias> defines a URL alias for matching HTTP requests. Any number of <host-alias> can be used for each alias.
The host-alias can be used either in the resin.xml or in a host.xml when use host-deploy together with resin:import.
Since the <host-deploy> and <host> tags lets you add a host.xml file to customize configuration, the <host-alias> can also fit in the custom host.xml page.
<host-alias-regexp> defines a regular expression for matching URLs for a given virtual host.
<host-default> configures defaults for a virtual host.
The host-default can contain any of the host configuration tags. It will be used as defaults for any virtual host.
<host-deploy> configures an automatic deployment directory for virtual host.
The following example configures /var/www/hosts as a host deployment directory. Each virtual host will have a webapps directory for .war deployment. So the directory /var/www/hosts/www.foo.com/webapps/bar/test.jsp would serve the URL http://www.foo.com/bar/test.jsp.
<host-name> defines the canonical name for a virtual host. The <host-name> will be used in Resin's logging, management, and is available in the host's variables.
All Resource tags are available to the <host>, for example, resources like <database> or <authenticator>. Resources defined at the host level are available for all web-apps in the host.
<rewrite-dispatch> defines a set of rewriting rules for dispatching and forwarding URLs. Applications can use these rules to redirect old URLs to their new replacements.
See rewrite-dispatch for more details.
<root-directory> configures the virtual host's filesystem root.
Because the virtual host's root will typically contain non-public files like log files, all web-apps should have a path below the host.
<secure-host-name> sets a host-name or URL to be used for secure redirection. For some security configurations, Resin needs to redirect from an insecure site to a secure one. The <secure-host-name> configures the host to redirect to.
See Resin security.
<web-app> configures a web application.
When specified by id, the application will be initialized on server start. When specified by url-regexp, the application will be initialized at the first request. This means that load-on-startup servlets may start later than expected for url-regexp applications.
The following example creates a web-app for /apache using the Apache htdocs directory to serve pages.
The following example sets the root web-app to the IIS root directory.
When the web-app is specified with a url-regexp, root-directory can use replacement variables ($2).
In the following, each user gets his or her own independent application using ~user.
<web-app-default> configures common values for all web applications.
Specifies war expansion.
web-app-deploy can be used in web-apps to define a subdirectory for war expansion. The tutorials in the documentation use web-app-deploy to allow servlet/tutorial/helloworld to be an independent war file.
The web-app-deploy can override configuration for an expanded war with a matching <web-app> inside the <web-app-deploy>. The <document-directory> is used to match web-apps.
The versioning attribute of the <web-app-deploy> tag improves web-app version updates by enabling a graceful update of sessions. The web-apps are named with numeric suffixes, e.g. foo-10, foo-11, etc, and can be browsed as /foo. When a new version of the web-app is deployed, Resin continues to send current session requests to the previous web-app. New sessions go to the new web-app version. So users will not be aware of the application upgrade.