The BCC resque bundle provides integration of php-resque to Symfony2. It is inspired from resque, a Redis-backed Ruby library for creating background jobs, placing them on multiple queues, and processing them later.
- Creating a Job, with container access in order to leverage your Symfony services
- Enqueue a Job with parameters on a given queue
- Creating background worker on a given queue
- A UX to monitor your queues, workers and job statuses
- ability to schedule jobs to run at a specific time or after a number of seconds delay
- ability to auto re-queue failed jobs, with back-off strategies
TODOs:
- Log management
- Job status tracking
- Redis configuration
- Localisation
- Tests
Make sure you have redis installed on your machine: http://redis.io/
Add to your bcc-resque-bundle
to your dependencies:
{
"require": {
...
"bcc/resque-bundle": "dev-master"
}
...
}
To install, run php composer.phar [update|install]
.
<?php
// app/AppKernel.php
public function registerBundles()
{
return array(
// ...
new BCC\ResqueBundle\BCCResqueBundle(),
// ...
);
}
Add to your routing.yml
:
# app/config/routing.yml
BCCResqueBundle:
resource: "@BCCResqueBundle/Resources/config/routing.xml"
prefix: /resque
You can customize the prefix as you wish.
You can now access the dashboard at this url: /resque
To secure the dashboard, you can add the following to your security.yml
- provided your administrator role is ROLE_ADMIN
access_control:
- { path: ^/resque, roles: ROLE_ADMIN }
Add to your security.yml
:
# app/config/security.yml
access_control:
- { path: ^/resque, roles: ROLE_ADMIN }
Now only users with the role ROLE_ADMIN will be able to access the dashboard at this url: /resque
You may want to add some configuration to your config.yml
# app/config/config.yml
bcc_resque:
class: BCC\ResqueBundle\Resque # the resque class if different from default
vendor_dir: %kernel.root_dir%/../vendor # the vendor dir if different from default
prefix: my-resque-prefix # optional prefix to separate Resque data per site/app
redis:
host: localhost # the redis host
port: 6379 # the redis port
database: 1 # the redis database
auto_retry: [0, 10, 60] # auto retry failed jobs
worker:
root_dir: path/to/worker/root # the root dir of app that run workers (optional)
See the Auto retry section for more on how to use auto_retry
.
Set worker: root_dir:
in case job fails to run when worker systems are hosted on separate server/dir from the system creating the queue.
When running multiple configured apps for multiple workers, all apps must be able to access by the same root_dir defined in worker: root_dir
.
This bundle is prepared for lazy loading in order to make a connection to redis only when its really used. Symfony2 supports that starting with 2.3. To make it work an additional step needs to be done. You need to install a proxy manager to your Symfony2 project. The full documentation for adding the proxy manager can be found in Symfony2's Lazy Service documentation.
A job is a subclass of the BCC\ResqueBundle\Job
class. You also can use the BCC\Resque\ContainerAwareJob
if you need to leverage the container during job execution.
You will be forced to implement the run method that will contain your job logic:
<?php
namespace My;
use BCC\ResqueBundle\Job;
class MyJob extends Job
{
public function run($args)
{
\file_put_contents($args['file'], $args['content']);
}
}
As you can see you get an $args parameter that is the array of arguments of your Job.
You can get the resque service simply by using the container. From your controller you can do:
<?php
// get resque
$resque = $this->get('bcc_resque.resque');
// create your job
$job = new MyJob();
$job->args = array(
'file' => '/tmp/file',
'content' => 'hello',
);
// enqueue your job
$resque->enqueue($job);
Executing the following commands will create a work on :
- the
default
queue :app/console bcc:resque:worker-start default
- the
q1
andq2
queue :app/console bcc:resque:worker-start q1,q2
(separate name with,
) - all existing queues :
app/console bcc:resque:worker-start "*"
You can also run a worker foreground by adding the --foreground
option;
By default VERBOSE
environment variable is set when calling php-resque
--verbose
option setsVVERBOSE
--quiet
disables both so no debug output is thrown
See php-resque logging option : https://github.com/chrisboulton/php-resque#logging
You can specify that a job is run at a specific time or after a specific delay (in seconds).
From your controller you can do:
<?php
// get resque
$resque = $this->get('bcc_resque.resque');
// create your job
$job = new MyJob();
$job->args = array(
'file' => '/tmp/file',
'content' => 'hello',
);
// enqueue your job to run at a specific \DateTime or int unix timestamp
$resque->enqueueAt(\DateTime|int $at, $job);
// or
// enqueue your job to run after a number of seconds
$resque->enqueueIn($seconds, $job);
You must also run a scheduledworker
, which is responsible for taking items out of the special delayed queue and putting
them into the originally specified queue.
app/console bcc:resque:scheduledworker-start
Stop it later with app/console bcc:resque:scheduledworker-stop
.
Note that when run in background mode it creates a PID file in 'cache//bcc_resque_scheduledworker.pid'. If you
clear your cache while the scheduledworker is running you won't be able to stop it with the scheduledworker-stop
command.
Alternatively, you can run the scheduledworker in the foreground with the --foreground
option.
Note also you should only ever have one scheduledworker running, and if the PID file already exists you will have to use
the --force
option to start a scheduledworker.
It's probably best to use supervisord (http://supervisord.org) to run the workers in production, rather than re-invent job spawning, monitoring, stopping and restarting.
Here's a sample conf file
[program:myapp_phpresque_default]
command = /usr/bin/php /home/sites/myapp/prod/current/vendor/bcc/resque-bundle/BCC/ResqueBundle/bin/resque
user = myusername
environment = APP_INCLUDE='/home/sites/myapp/prod/current/vendor/autoload.php',VERBOSE='1',QUEUE='default'
stopsignal=QUIT
[program:myapp_phpresque_scheduledworker]
command = /usr/bin/php /home/sites/myapp/prod/current/vendor/bcc/resque-bundle/BCC/ResqueBundle/bin/resque-scheduler
user = myusername
environment = APP_INCLUDE='/home/sites/myapp/prod/current/vendor/autoload.php',VERBOSE='1',RESQUE_PHP='/home/sites/myapp/prod/current/vendor/chrisboulton/php-resque/lib/Resque.php'
stopsignal=QUIT
[group:myapp]
programs=myapp_phpresque_default,myapp_phpresque_scheduledworker
(If you use a custom Resque prefix, add an extra environment variable: PREFIX='my-resque-prefix')
Then in Capifony you can do
sudo supervisorctl stop myapp:*
before deploying your app and sudo supervisorctl start myapp:*
afterwards.
You can change a job queue just by setting the queue
field of the job:
From within the job:
<?php
namespace My;
use BCC\ResqueBundle\Job;
class MyJob extends Job
{
public function __construct()
{
$this->queue = 'my_queue';
}
public function run($args)
{
...
}
}
Or from outside the job:
<?php
// create your job
$job = new MyJob();
$job->queue = 'my_queue';
Just extend the ContainerAwareJob
:
<?php
namespace My;
use BCC\ResqueBundle\ContainerAwareJob;
class MyJob extends ContainerAwareJob
{
public function run($args)
{
$doctrine = $this->getContainer()->getDoctrine();
...
}
}
Use the app/console bcc:resque:worker-stop
command.
- No argument will display running workers that you can stop.
- Add a worker id to stop it:
app/console bcc:resque:worker-stop ubuntu:3949:default
- Add the
--all
option to stop all the workers.
You can have the bundle auto retry failed jobs by adding retry strategy
for either a specific job, or for all jobs in general:
The following will allow Some\Job
to retry 3 times.
- right away
- after a 10 second delay
- after a 60 second delay
bcc_resque:
redis:
....
auto_retry:
Some\Job: [0, 10, 60]
Setting strategy for all jobs:
bcc_resque:
auto_retry: [0, 10, 60]
With default strategy for all but specific jobs:
bcc_resque:
auto_retry:
default: [0, 10, 60]
Some\Job: [0, 10, 120, 240]
Some\Other\Job: [10, 30, 120, 600]
The default
strategy (if provided) will be applied to all jobs that does not have a specific strategy attached. If not provided these jobs will not have auto retry.
You can disable auto_retry
for selected jobs by using an empty array:
bcc_resque:
auto_retry:
default: [0, 10, 60]
Some\Job: []
Some\Other\Job: [10, 30, 120, 600]
Here Some\Job
will not have any auto_retry
attached.
Please note
To use the auto_retry
feature, you must also run the scheduler job:
app/console bcc:resque:scheduledworker-start