Delayed Job

Delayed_job (or DJ) encapsulates the common pattern of asynchronously executing longer tasks in the background.

This is a direct port of Delayed::Job and this fork also supports MySQL.

The reason I've added MySQL support is because I required a data store that will ensure transactions for the project I plan to embed this into and I don't want to maintain an mongo installation just for the sake of this plugin.

This can be modified to use any Database adapter or data source that you want but so far I've only tested MySQL and the model will throw an exception if you try to use something other than mongo and MySQL. Feel free to fork and send me a pull request.

Pre-Requisites - If using MongoDB

• Working installation of MongoDB

  http://www.mongodb.org/display/DOCS/Quickstart+Unix

• Working installation of Mongo PEAR extension

  apt-get install php5-dev php-pear

  pecl install mongo

  echo 'extension=mongo.so' > /etc/php5/apache2/conf.d/mongo.ini

  service apache2 restart

Pre-Requisites - If using MySQL

• Working installation/connection to a MySQL server

• You can safely comment use MongoDate; in li3_delayed_jobs\models\Jobs.php

In fact, you must comment this line unless the Mongo PEAR extension is enabled in your PHP installation. (Just do it if not using mongo for this).

Installation

Check out the code to your library directory

cd libraries
git clone [email protected]:cgarvis/li3_delayed_job.git

Include the library in your /app/config/bootstrap/libraries.php

Libraries::add('li3_delayed_job');

Configuration

• In the li3_delayed_job\models\Jobs.php you can modify the following:
• Jobs::storeObject determines whether or not to keep the queue objects in the table
• Jobs::destroyFailedJobs determines whether or not to erase failed tasks after attempts are exhausted
• Jobs::keyID Stores the default storage engine table/document table primary key. Currently set automatically to _id for mongo and id for database (assuming MySQL). Look at Jobs::setDataSourceType() for change these.

Usage

Jobs are simple objects with a method called perform. Any object which responds to perform can be stuffed into the jobs collection. Job objects are serialized to yaml so that they can later be resurrected by the job runner.

Creating a job

See examples under ./tests.

Running the jobs

You can invoke li3 jobs work which will start working off jobs. You can cancel the task with CTRL-C

Keep in mind that each worker will check the database at least every 5 seconds.

Cleaning up

You can invoke li3 jobs clear to delete all jobs in the queue