Setup prototype for Liferay development with Vagrant

This project contains the Vagrant project that creates a VM, provisioned with a Liferay server and it's prerequisites. Provisioning is performed using an Ansible playbook. It installs:

• a MySQL database
• a Java 7 JDK
• a Liferay 6.2 Tomcat bundle.

The packages for MySQL and Java are installed using Aptitude. Local caching for (many of) those packages is prepared. It is activated when the Vagrant plugin vagrant-cachier is installed.

The packages for the Liferay system are downloaded from configurable locations. They are cached in the directory .ansible/cache/download.

The auto deploy directory of the Liferay server is accessible from outside the VM, so that each (developed) Liferay extension can be deployed only by copying its *.war archive to that directory.

Prerequisites

Starting and provisioning the VM requires the following software packages to be installed:

• Vagrant
• Ansible

Starting and using the VM

See Vagrant Documentation how to work with Vagrant. Important commands are:

vagrant up
vagrant provision
vagrant halt
vagrant destroy

Using the VM

After the VM is successfully started and provisioned, open http://localhost:18080/ in a browser and follow the requested steps. Sign in with the default admin user:

• Email Address: test@liferay.com
• Password: test

Extension development

The extension development might e.g. be organised in a Maven (multi module) project. It should be configured with the path to the auto deploy directory.

See Setup prototype for Liferay development with Maven.

Server setup with Ansible

The Ansible playbook, that is used to provision the Vagrant VM, can also be used to set up the Liferay system (including MySQL and Java) on other environments (servers), too. For each additional environment, an inventory file provisioning/inventory_[environment] has to be created. Replicate an existing inventory file and adjust its contents regarding to the new environment and/or see the documentation on the Ansible home page: Ansible Docs - Inventory. If the new environment requires custom configurations, create a file provisioning/host_vars/[environment].yml and configure the variables that should be overwritten.

To run the setup call:

ansible-playbook provisioning/playbook.yml \
                 -i provisioning/inventory_[environment]

Variables for Liferay version

The default variables define the (currently) latest version of the Liferay Community Edition for installation. If a server should be set up with another edition/version, those variables can be overwritten, e.g. in the regarding provisioning/host_vars/[environment].yml file for the server to be set up. Prefedined variables for known editions can be found in the directory provisioning/roles/liferay/edition_vars and can simply be copied to the provisioning/host_vars/[environment].yml file.

TODOs / Improvements

TODOs and improvments for this setup:

• The file provisioning/roles/mysql/tasks/main.yml defines a task ensure root access, that provides access rights to the root user, so that the database and the database user can be for the Liferay system can be created by the Ansible scripts. The current solution for this task is good enough to set up a Vagrant VM, but it shouldn't be used to set up a productive or any other system, that is accessible from the internet.
• The setup of the Tomcat service is not properly implemented, it normally should work with daemon.sh and jsvc. That setup caused many problems and we decided to do a first attempt with a custom service.sh script, that simply invokes Tomcat's catalina.sh. It's not 100% stable but basically seems to work. Anyway, we should rather implement a proper solution with daemon.sh and jsvc.
• Avoid Liferay installation if {{liferay_home_dir}} already exists.
• Add more provisioning/roles/liferay/edition_vars/liferay-[VERSION].yml files for further known Liferay editions.

Pull requests addressing any of those TODOs are very welcome!