Getting Started: Deploying a Web Application

Welcome back! This article covers creating a basic Ring web application and deploying to an Immutant. It is the second installment in a series of tutorials on getting started with Immutant. If you haven't read the first installment, go do so now, since it covers installation and setup. This tutorial assumes you are on a *nix system.

Creating an Immutant Clojure application

In our previous article, we installed the lein plugin. Let's take another look at the tasks it provides:

~/immutant $ lein immutant
Manage the deployment lifecycle of an Immutant application.

Subtasks available:
new        Create a new project skeleton initialized for immutant.
init       Adds a sample immutant.clj configuration file to an existing project
deploy     Deploys the current project to the Immutant specified by $IMMUTANT_HOME
undeploy   Undeploys the current project from the Immutant specified by $IMMUTANT_HOME
run        Starts up the Immutant specified by $IMMUTANT_HOME, displaying its console output

We talked about run last time. This time, we'll cover new and deploy. To do so, we'll build a basic application that demonstrates the current web features. To get started, let's create an Immutant project:

~/immutant $ lein immutant new immutant-demo
Created new project in: /Users/tobias/immutant/immutant-demo
Look over project.clj and start coding in immutant_demo/core.clj
Wrote sample immutant.clj

The new task creates a Leiningen project and gives it a sample Immutant configuration file (immutant.clj). It is equivalent to calling:

~/immutant $ lein new immutant-demo && cd immutant-demo && lein immutant init

We'll come back to immutant.clj in a sec. Now, let's add a ring handler to our core namespace:

(ns immutant-demo.core)

(defn ring-handler [request]
  {:status 200
    :headers {"Content-Type" "text/html"}
    :body "Hello from Immutant!" })

Configuring the application for Immutant

When the Immutant deploys an application, it looks for a file named immutant.clj at the root and evaluates it if it exists. This file is used to configure the Immutant services you want your application to consume. It's the single place you defines all the components required by your application, and saves you from having to keep external configuration files in sync (crontabs, message queue definitions, init scripts, etc).

The file has example code for configuring web endpoints and messaging services, but we're just going to deal with web endpoints in this article. Edit your immutant.clj so it looks like:

(ns immutant-demo.init
  (:use immutant-demo.core)
  (:require [immutant.messaging :as messaging]
            [immutant.web :as web]))

(web/start "/" #'ring-handler)

We'll come back to what web/start is doing after we get the application running.

Deploying your application

Before we can start up an Immutant, we need to tell it about our application. We do that by deploying (for this to work, you need to have IMMUTANT_HOME set - see the previous article for details):

~/immutant/immutant-demo $ lein immutant deploy
Deployed immutant-demo to /Users/tobias/immutant/current/jboss/standalone/deployments/immutant-demo.clj

This writes a deployment descriptor to Immutant's deploy directory which points back to the application's root directory. Now the Immutant can find your application - so let's fire it up.

Starting Immutant

To launch an Immutant, use the lein immutant run command. This will start the Immutant's JBoss server, and will run in the foreground displaying the console log. You'll see lots of log messages that you can ignore - the one to look for should be the last message, and should tell you the app was deployed:

~/immutant/immutant-demo $ lein immutant run
Starting Immutant via /Users/tobias/immutant/current/jboss/bin/standalone.sh
...
(a plethora of log messages deleted)
...
13:04:39,888 INFO  [org.jboss.as.server.controller] (DeploymentScanner-threads - 2) Deployed "immutant-demo.clj"

Now, let's verify that our app is really there. Immutant runs on port 8080 by default, so let's hit it and see what happens:

~ $ curl http://localhost:8080/immutant-demo/
Hello from Immutant!

Yay!

You can kill the Immutant with Ctrl-C.

Context Paths

Remember our call to web/start earlier? Let's talk about what that is doing. To do that, however, we need to first talk about context paths. The context path is the portion of the URL between the hostname and the routes within the application. It basically tells Immutant which requests to route to a particular application.

An Immutant can host multiple applications at the same time, but each application must have a unique context path. If no context path is provided when an application is deployed, it defaults to one based on the name of the deployment. The deployment name is taken from the name of the deployment descriptor, which in turn is taken from the name of the project given to defproject in project.clj. So for our sample app above, the context path defaults to /immutant-demo. You can override this default by specifying a :context-path within an :immutant map in your project.clj. Let's go ahead and do that:

(defproject immutant-demo "1.0.0-SNAPSHOT"
  :description "A basic demo"
  :dependencies [[org.clojure/clojure "1.2.1"]]
  :immutant {:context-path "/"})

Now, when you call lein immutant deploy, the context path will be picked up from your project.clj and included in the deployment descriptor and any web endpoints your application stands up will be accessible under that context path.

Which brings us back to web/start. web/start stands up a web endpoint for you, and takes two arguments: a sub-context path and a Ring handler function. The sub-context path is relative to the application's context path, so a context path of "/ham" and a sub-context path of "/" makes the handler function available at /ham, whereas a sub-context path of "/biscuits" makes the handler function available at /ham/biscuits. Make sense?

You can register as many web endpoints as you like within an application - they just each need an application unique sub-context path. If we add this to our core.clj:

(defn another-ring-handler [request]
  {:status 200
   :headers {"Content-Type" "text/html"}
   :body "Pssst! Over here!"})

And this to our immutant.clj:

(web/start "/biscuits" #'another-ring-handler)

Redeploy the application to pick up the :context-path from immutant.clj:

~/immutant/immutant-demo $ lein immutant deploy
Deployed immutant-demo to /Users/tobias/immutant/current/jboss/standalone/deployments/immutant-demo.clj

Then fire an Immutant up again with lein immutant run, we can see they both work:

~ $ curl http://localhost:8080
Hello from Immutant!
~ $ curl http://localhost:8080/biscuits
Pssst! Over here!

web/start has a companion function for shutting down a web endpoint: web/stop. It takes the sub-context path for the endpoint, and can be called from anywhere. You aren't required to shut down your endpoints - Immutant will do that on your behalf when it is shut down or the application is undeployed.

Wrapping up

I hope you've enjoyed this quick run-through of deploying a web application to Immutant. Since Immutant is still in a pre-alpha state, none of what I said above is set in stone. If anything does change, I'll edit this post to keep it accurate. I've posted the demo application we've built if you want to download it.

If you have any feedback or questions, get in touch! And stay tuned - our next tutorial will cover using Immutant's messaging features.