Resque for PHP
About
This is a fork of the vent/resque repo on https://github.com/vend/php-resque. The idea is to improve on the package without the need to rely on implementations of vent. It seems the development on the original bundle (chrisboulton/php-resque) and vend stopped a long time ago. The namespacing seems to be a huge improvement over the original implementation. Future updates will include dequeueing and more., (*1)
Namespaced Fork
, (*2)
Resque is a Redis-backed library for creating background jobs, placing
those jobs on one or more queues, and processing them later., (*3)
This is a PHP fork of the Resque worker and job classes. This makes it compatible with the
resque-web interface, and with other Resque libraries. (You could enqueue jobs from Ruby and dequeue
them in PHP, for instance)., (*4)
This library (vend/resque) is a fork of chrisboulton/php-resque at around
version 1.3, that has been refactored to remove global state, add namespacing, and improve
decoupling. This makes it easier to use and extend., (*5)
Getting Started
Add vend/resque to your application's composer.json., (*6)
{
"require": {
"vend/resque": "~2.1.0"
}
}
Requirements
- PHP 5.3+
- A Redis client library (for instance, Predis or Credis)
Jobs
Queueing Jobs
Jobs are queued as follows:, (*7)
use Resque\Resque;
use Predis\Client;
$resque = new Resque(new Client());
$resque->enqueue('default_queue', 'App\Job', array('foo' => 'bar'), true);
In order the arguments are: queue, job class, job payload, whether to enable tracking., (*8)
Defining Jobs
Each job should be in its own class, and implement the Resque\JobInterface. (This is pretty easy,
and only really requires a single custom method: perform().) Most of the time, you'll want to
use the default implementation of a Job, and extend the Resque\AbstractJob instead of implementing
the interface yourself:, (*9)
namespace App;
use Resque\AbstractJob;
class Job extends AbstractJob
{
public function perform()
{
// work work work
$this->doSomething($this->payload['foo']);
}
}
Any exception thrown by a job will result in the job failing - be
careful here and make sure you handle the exceptions that shouldn't
result in a job failing., (*10)
Job Status Tracking
vend/resque has the ability to perform basic status tracking of a queued
job. The status information will allow you to check if a job is in the
queue, is currently being run, has finished, or has failed., (*11)
To track the status of a job, pass true as the fourth argument to
Resque\Resque::enqueue. An ID used for tracking the job status will be
returned:, (*12)
$id = $resque->enqueue('default_queue', 'App\Job', $payload, true);
echo $id; // [0-9a-f]{32}
To fetch the status of a job:, (*13)
$factory = new Resque\Job\StatusFactory($resque);
// Pass the ID returned from enqueue
$status = $factory->forId($id);
// Alternatively, to get the status for a Job instance:
$status = $factory->forJob($job);
// Outputs the status as a string: 'waiting', 'running', 'complete', etc.
echo $status->getStatus();
The Status object contains methods for adding other attributes to the
tracked status. (For instance, you might use the status object to track
errors, completion information, iterations, etc.), (*14)
Statuses
Job statuses are defined as constants in the Resque\Job\Status class.
Valid statuses include:, (*15)
-
Resque\Job\Status::STATUS_WAITING - Job is still queued
-
Resque\Job\Status::STATUS_RUNNING - Job is currently running
-
Resque\Job\Status::STATUS_FAILED - Job has failed
-
Resque\Job\Status::STATUS_COMPLETE - Job is complete
Statuses are available for up to 24 hours after a job has completed
or failed, and are then automatically expired. A status can also
forcefully be expired by calling the stop() method on a status
class., (*16)
Console
You are free you implement your own daemonization/worker pool strategy by
subclassing the Worker class. For playing around, and low throughput
applications, you might like to use the built-in console commands., (*17)
This version of the library uses the Symfony2 Console component. But it
must be configured with the details of your Redis connection. We do this
in much the same way that the Doctrine2 Console component gets details
of your database connection: via a cli-config.php file., (*18)
There is an example cli-config.php in this repository., (*19)
If you're running a full-stack web application, you'd generally use your
locator/service container to fill in the Redis client connection in this
file. (The default is to use Predis, and to connect to 127.0.0.1:6379)., (*20)
Basic Usage
resque <subcommand>
Available commands:
enqueue Enqueues a job into a queue
help Displays help for a command
list Lists commands
worker Runs a Resque worker
queue
queue:clear Clears a specified queue
queue:list Outputs information about queues
Enqueueing
This command will enqueue a Some\Test\Job job onto the default queue.
Watch out for the single quotes around the class name: when specifying backslashes
on the command line, you'll probably have to avoid your shell escaping them., (*21)
resque enqueue default 'Some\Test\Job' -t
Worker
This command will run a simple pre-forking worker on two queues:, (*22)
resque worker -Q default -Q some_other_queue
(-q means quiet, -Q specifies queues). You can also specify no queues, or
the special queue '*' (watch for shell expansion)., (*23)
There are a couple of useful commands for getting information about the queues. This
will show a list of queues and how many jobs are waiting on each:, (*24)
resque queue:list
This command will clear a specified queue:, (*25)
resque queue:clear default
Logging
The library now uses PSR3. When running as a console component, you can customise
the logger to use in cli-config.php. (For instance, you might like to send your
worker logs to Monolog.), (*26)
Why Fork?
Unfortunately, several things about the existing versions of php-resque made it
a candidate for refactoring:, (*27)
- php-resque supported the Credis connection library and had hard-coded this
support by using static accessors. This meant more advanced features such as
replication and pipelining were unavailable.
- Now, Resque for PHP supports any client object that implements a suitable
subset of Redis commands. No type-checking is done on the passed in connection,
meaning you're free to use Predis, or whatever you like.
- While the public API of
php-resque was alright, the protected API was pretty
much useless. This made it hard to extend worker classes to dispatch jobs differently.
- Important state (the underlying connection) was thrown into the global scope
by hiding it behind static accessors (
Resque::redis()). This makes things
easier for the library author (because he/she need not think about dependencies)
but also statically ties together the classes in the library: it makes
testing and extending the library hard.
- There's no reason to do this:
Resque instances should simply use DI and
take a client connection as a required constructor argument.
- This improvement also allows the connection to be mocked without extending
the
Resque class.
- And it lets you reuse your existing connection to Redis, if you have one.
No need to open a new connection just to enqueue a job.
- Statistic classes were static for no good reason.
- To work, statistics need a connection to Redis, a name, and several methods
(get, set). State and methods to manipulate it? Sounds like a task for
objects! Not just static methods, that don't encapsulate any of the state.
- Because these were static calls, the
Resque_Stat class was hard-coded into
several other classes, and could not be easily extended.
- The library is now fully namespaced and compatible with PSR-0. The top level
namespace is
Resque.
- The events system has been removed. There is now little need for it.
It seems like the events system was just a workaround due to the poor
extensibility of the Worker class. The library should allow you to extend any
class you like, and no method should be too long or arduous to move into a
subclass.
Contributors
Here's the contributor list from earlier versions, at chrisboulton/php-resque:, (*28)
- @chrisboulton
- @acinader
- @ajbonner
- @andrewjshults
- @atorres757
- @benjisg
- @cballou
- @chaitanyakuber
- @charly22
- @CyrilMazur
- @d11wtq
- @danhunsaker
- @dceballos
- @ebernhardson
- @hlegius
- @hobodave
- @humancopy
- @JesseObrien
- @jjfrey
- @jmathai
- @joshhawthorne
- @KevBurnsJr
- @lboynton
- @maetl
- @matteosister
- @MattHeath
- @mickhrmweb
- @Olden
- @patrickbajao
- @pedroarnal
- @ptrofimov
- @rajibahmed
- @richardkmiller
- @Rockstar04
- @ruudk
- @salimane
- @scragg0x
- @scraton
- @thedotedge
- @tonypiper
- @trimbletodd
- @warezthebeef
Wallogit.com
