Play 2.1 integration with Cloud Foundry

Is there anybody having the problems with integration of Play 2.1 applications with Cloud Foundry?

If so, please check this post out, since its author has just spent about 3h trying to integrate the latest (so far) 2.1 version of Play framework with Cloud Foundry PaaS, and in the end all efforts have been paid off to him.

Well, everything had started when I wanted to move from Anorm to Slick for DB management and that's why I needed Scala 2.10.0 and thus Play 2.1.
I already knew for integration problems of Play 2.1 and Cloud Foundry, but on the other hand CF is one of the promising PaaS platforms which offers great resources in its beta phase (2GB RAM, arbitrary number of applications instances, PostgreSQL 9.0, RabitMQ, and few other services - all for free), so I just didn't want to miss a thing.

On the many places I could read how "Cloud Foundry doesn't support Play 2.1", how "it breaks its integration", etc., etc., and guess what - I've just managed to solve this issue and now my application is running on Play 2.1 right there: http://pijaca.cloudfoundry.com/ (okay, I'm still developing it, so please, leave critics behind for app itself ;)).

Since I guessed that there could be also people who want to try Play 2.1 with Cloud Foundry for various reasons, I just wanted to explain how I solved my integration issues, in the following steps.

Play 2.1 with Cloud Foundry - Integration Steps

1. Create Play 2.1 application or migrate your existing Play 2.0.x applications to Play 2.1. If you are doing migration, take special care to project/Build.scala, project/plugins.sbt and project/build.properties files.

In my case, they look like this:

Build.scala


import sbt._
import Keys._
import play.Project._

object ApplicationBuild extends Build {

  val appName         = "pijaca"
  val appVersion      = "1.0-SNAPSHOT"

  val appDependencies = Seq(
    // Add your project dependencies here,
    jdbc,
    anorm,
    "postgresql" % "postgresql" % "8.4-702.jdbc4",
    "org.hibernate" % "hibernate-entitymanager" % "3.6.9.Final"
  )

  val main = play.Project(appName, appVersion, appDependencies).settings(
    // Add your own project settings here      
  )
}

plugins.sbt


// Comment to get more information during initialization
logLevel := Level.Warn

// The Typesafe repository
resolvers += "Typesafe repository" at "http://repo.typesafe.com/typesafe/releases/"

// Use the Play sbt plugin for Play projects
addSbtPlugin("play" % "sbt-plugin" % "2.1.0")


build.properties

sbt.version=0.12.2


2. Download the next libraries:

auto-reconfiguration-0.6.5.jar
play-java-jpa_2.10-2.1.0.jar
hibernate-jpa-2.0-api-1.0.1.Final.jar

from the next locations (respectively):

http://rapidshare.com/files/2061661742/auto-reconfiguration-0.6.5.jar
(original link to vcap-staging project containing built library became inactive in mid-time so I have uploaded it onto Rapidshare)

http://www.versioneye.com/package/play--play-java-jpa_2~10/version/2~1~0

http://www.versioneye.com/package/org~hibernate~javax~persistence--hibernate-jpa-2~0-api/version/1~0~1~Final

Just to mention that here also figures play-java-jdbc_2.10_2.1.0.jar as an additional dependency, but for me it worked even without this library, probably because on my classpath I already had play-java_xxx.jar and play-jdbc_xxx.jar. Since I didn't have problem with this one I haven't checked this in more details.

3. Add the mentioned three libraries onto your classpath.

4. In your conf folder create cloud.conf and cloudfoundry.properties files. They should look like:

cloud.conf

include "application.conf"
include "cloudfoundry.properties"

db.default.driver=${?cloud.services.postgresql.connection.driver}
db.default.url=${?cloud.services.postgresql.connection.url}
db.default.password=${?cloud.services.postgresql.connection.password}
db.default.user=${?cloud.services.postgresql.connection.username}

cloudfoundry.properties

autoconfig=false

Of course, application.conf is your old configuration file. cloud.conf file from this example overrides settings only for PostgreSQL database. In my case, I have only one instance of this service, so that's why "postgresql" figures in those settings. Otherwise, instead of it you should place the name of your service, e.g. "postgresql_service_1".


5. Clean your project. (Because of moving from different versions of Play and Scala I performed first "play clean-all").

play clean compile
play -Dconfig.file=path-to-conf\cloud.conf dist 

and since I already saved my deploying information in manifest.yml file, I have just executed:

vmc push

to deploy my application to the Cloud Foundry.

For the curious ones, my manifest.yml file looks something like this:

---
applications:
- name: pijaca
  framework: play
  runtime: java7
  memory: 1G
  instances: 1
  url: pijaca.${target-base}
  path: dist/pijaca-1.0-SNAPSHOT.zip
  services:
    PostgreSQL_1:
      vendor: postgresql
      version: '9.0'
      tier: free

After couple of minutes my application is uploaded and now it's running happily on CF! :)

So far, that should be all you need to deploy your Play 2.1 application onto the Cloud Foundry. If I found something else that I missed here, I'll update this post later on.

Hopefully, someone will find this post helpful in the end. 

Cheers,

Alex