Module: Resque

Extended by:
Resque
Includes:
Helpers
Included in:
Resque
Defined in:
lib/resque.rb,
lib/resque/job.rb,
lib/resque/stat.rb,
lib/resque/errors.rb,
lib/resque/plugin.rb,
lib/resque/server.rb,
lib/resque/worker.rb,
lib/resque/delayed.rb,
lib/resque/failure.rb,
lib/resque/helpers.rb,
lib/resque/version.rb,
lib/resque/failure/base.rb,
lib/resque/failure/mongo.rb,
lib/resque/failure/redis.rb,
lib/resque/failure/airbrake.rb,
lib/resque/failure/multiple.rb,
lib/resque/server/test_helper.rb

Defined Under Namespace

Modules: Delayed, DelayedJob, Failure, Helpers, Plugin, Stat, TestHelper Classes: DelayedQueueError, DirtyExit, Job, NoClassError, NoQueueError, QueueError, Server, Worker

Constant Summary collapse

Version =
VERSION = '1.20.0'

Instance Method Summary collapse

Methods included from Helpers

#classify, #constantize, #decode, #encode

Instance Method Details

#after_fork(&block) ⇒ Object

The ‘after_fork` hook will be run in the child process and is passed the current job. Any changes you make, therefore, will only live as long as the job currently being processed.

Call with a block to set the hook. Call with no arguments to return the hook.



96
97
98
# File 'lib/resque.rb', line 96

def after_fork(&block)
  block ? (@after_fork = block) : @after_fork
end

#after_fork=(after_fork) ⇒ Object

Set the after_fork proc.



101
102
103
# File 'lib/resque.rb', line 101

def after_fork=(after_fork)
  @after_fork = after_fork
end

#before_first_fork(&block) ⇒ Object

The ‘before_first_fork` hook will be run in the parent process only once, before forking to run the first job. Be careful- any changes you make will be permanent for the lifespan of the worker.

Call with a block to set the hook. Call with no arguments to return the hook.



65
66
67
# File 'lib/resque.rb', line 65

def before_first_fork(&block)
  block ? (@before_first_fork = block) : @before_first_fork
end

#before_first_fork=(before_first_fork) ⇒ Object

Set a proc that will be called in the parent process before the worker forks for the first time.



71
72
73
# File 'lib/resque.rb', line 71

def before_first_fork=(before_first_fork)
  @before_first_fork = before_first_fork
end

#before_fork(&block) ⇒ Object

The ‘before_fork` hook will be run in the parent process before every job, so be careful- any changes you make will be permanent for the lifespan of the worker.

Call with a block to set the hook. Call with no arguments to return the hook.



81
82
83
# File 'lib/resque.rb', line 81

def before_fork(&block)
  block ? (@before_fork = block) : @before_fork
end

#before_fork=(before_fork) ⇒ Object

Set the before_fork proc.



86
87
88
# File 'lib/resque.rb', line 86

def before_fork=(before_fork)
  @before_fork = before_fork
end

#collection_for_queue(queue) ⇒ Object

Returns the mongo collection for a given queue



247
248
249
250
# File 'lib/resque.rb', line 247

def collection_for_queue(queue)
  queue = namespace_queue(queue)
  mongo[queue]
end

#delayed_job?(klass) ⇒ Boolean

Returns:

  • (Boolean)


110
111
112
113
# File 'lib/resque.rb', line 110

def delayed_job?(klass)
  klass.instance_variable_get(:@delayed) ||
    (klass.respond_to?(:delayed) and klass.delayed)
end

#delayed_queue?(queue) ⇒ Boolean

Returns:

  • (Boolean)


115
116
117
# File 'lib/resque.rb', line 115

def delayed_queue?(queue)
  @delayed_queues.include? namespace_queue(queue)
end

#delayed_size(queue) ⇒ Object



185
186
187
188
189
190
191
192
# File 'lib/resque.rb', line 185

def delayed_size(queue)
  queue = namespace_queue(queue)
  if delayed_queue? queue
    mongo[queue].find({'delay_until' => { '$gt' => Time.now }}).count
  else
    mongo[queue].count
  end
end

#dequeue(klass, *args) ⇒ Object

This method can be used to conveniently remove a job from a queue. It assumes the class you’re passing it is a real Ruby class (not a string or reference) which either:

a) has a @queue ivar set
b) responds to `queue`

If either of those conditions are met, it will use the value obtained from performing one of the above operations to determine the queue.

If no queue can be inferred this method will raise a ‘Resque::NoQueueError`

If no args are given, this method will dequeue all jobs matching the provided class. See ‘Resque::Job.destroy` for more information.

Returns the number of jobs destroyed.

Example:

# Removes all jobs of class `UpdateNetworkGraph`
Resque.dequeue(GitHub::Jobs::UpdateNetworkGraph)

# Removes all jobs of class `UpdateNetworkGraph` with matching args.
Resque.dequeue(GitHub::Jobs::UpdateNetworkGraph, 'repo:135325')

This method is considered part of the ‘stable` API.



338
339
340
341
342
343
344
345
346
347
348
349
350
# File 'lib/resque.rb', line 338

def dequeue(klass, *args)
  # Perform before_dequeue hooks. Don't perform dequeue if any hook returns false
  before_hooks = Plugin.before_dequeue_hooks(klass).collect do |hook|
    klass.send(hook, *args)
  end
  return if before_hooks.any? { |result| result == false }

  Job.destroy(queue_from_class(klass), klass, *args)

  Plugin.after_dequeue_hooks(klass).each do |hook|
    klass.send(hook, *args)
  end
end

#dropObject



431
432
433
434
# File 'lib/resque.rb', line 431

def drop
  mongo.collections.each{ |collection| collection.drop unless collection.name =~ /^system./ }
  @mongo = nil
end

#enable_delay(queue) ⇒ Object



119
120
121
122
123
124
125
# File 'lib/resque.rb', line 119

def enable_delay(queue)
  queue = namespace_queue(queue)
  unless delayed_queue? queue
    @delayed_queues << queue
    mongo_stats.update({:stat => 'Delayed Queues'}, {'$addToSet' => {'value' => queue}}, {:upsert => true})
  end
end

#enqueue(klass, *args) ⇒ Object

This method can be used to conveniently add a job to a queue. It assumes the class you’re passing it is a real Ruby class (not a string or reference) which either:

a) has a @queue ivar set
b) responds to `queue`

If either of those conditions are met, it will use the value obtained from performing one of the above operations to determine the queue.

If no queue can be inferred this method will raise a ‘Resque::NoQueueError`

Returns true if the job was queued, nil if the job was rejected by a before_enqueue hook.

This method is considered part of the ‘stable` API.



278
279
280
# File 'lib/resque.rb', line 278

def enqueue(klass, *args)
  enqueue_to(queue_from_class(klass), klass, *args)
end

#enqueue_delayed(klass, *args) ⇒ Object



307
308
309
# File 'lib/resque.rb', line 307

def enqueue_delayed(klass, *args)
  
end

#enqueue_to(queue, klass, *args) ⇒ Object

Just like ‘enqueue` but allows you to specify the queue you want to use. Runs hooks.

‘queue` should be the String name of the queue you’re targeting.

Returns true if the job was queued, nil if the job was rejected by a before_enqueue hook.

This method is considered part of the ‘stable` API.



291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
# File 'lib/resque.rb', line 291

def enqueue_to(queue, klass, *args)
  # Perform before_enqueue hooks. Don't perform enqueue if any hook returns false
  before_hooks = Plugin.before_enqueue_hooks(klass).collect do |hook|
    klass.send(hook, *args)
  end
  return nil if before_hooks.any? { |result| result == false }

  Job.create(queue, klass, *args)

  Plugin.after_enqueue_hooks(klass).each do |hook|
    klass.send(hook, *args)
  end

  return true
end

#infoObject

Returns a hash, similar to redis-rb’s #info, of interesting stats.



412
413
414
415
416
417
418
419
420
421
422
423
# File 'lib/resque.rb', line 412

def info
  return {
    :pending   => queues.inject(0) { |m,k| m + size(k) },
    :processed => Stat[:processed],
    :queues    => queues.size,
    :workers   => workers.size.to_i,
    :working   => working.count,
    :failed    => Stat[:failed],
    :servers   => to_s,
    :environment  => ENV['RAILS_ENV'] || ENV['RACK_ENV'] || 'development'
  }
end

#initialize_mongoObject



39
40
41
42
43
44
# File 'lib/resque.rb', line 39

def initialize_mongo
  mongo_workers.create_index :worker
  mongo_stats.create_index :stat
  delayed_queues = mongo_stats.find_one(:stat => 'Delayed Queues')
  @delayed_queues = delayed_queues['value'] if delayed_queues
end

#inline=(inline) ⇒ Object



135
136
137
# File 'lib/resque.rb', line 135

def inline=(inline)
  @inline = inline
end

#inline?Boolean Also known as: inline

If ‘inline’ is true Resque will call #perform method inline without queuing it into Redis and without any Resque callbacks. The ‘inline’ is false Resque jobs will be put in queue regularly.

Returns:

  • (Boolean)


130
131
132
# File 'lib/resque.rb', line 130

def inline?
  @inline
end

#keysObject

Returns an array of all known Resque keys in Redis. Redis’ KEYS operation is O(N) for the keyspace, so be careful - this can be slow for big databases.



427
428
429
# File 'lib/resque.rb', line 427

def keys
  names = mongo.collection_names
end

#list_range(key, start = 0, count = 1, mode = :ready) ⇒ Object

Does the dirty work of fetching a range of items from a Redis list and converting them into Ruby objects.



218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
# File 'lib/resque.rb', line 218

def list_range(key, start = 0, count = 1, mode = :ready)
  query = { }
  sort = []
  if delayed_queue? key
    if mode == :ready
      query['delay_until'] = { '$not' => { '$gt' => Time.new}}
    elsif mode == :delayed
      query['delay_until'] = { '$gt' => Time.new}
    elsif mode == :delayed_sorted
      query['delay_until'] = { '$gt' => Time.new}
      sort << ['delay_until', 1]
    elsif mode == :all_sorted
      query = {}
      sort << ['delay_until', 1]
    end
  end
  queue = namespace_queue(key)
  items = mongo[queue].find(query, { :limit => count, :skip => start, :sort => sort}).to_a.map{ |i| i}
  count > 1 ? items : items.first
end

#mongoObject

Returns the current Mongo::DB. If none has been created, it will create a new one called ‘resque’.



33
34
35
36
37
# File 'lib/resque.rb', line 33

def mongo
  return @mongo if @mongo
  self.mongo = Mongo::Connection.new.db("resque")
  @mongo
end

#mongo=(database) ⇒ Object

Set the queue database. Expects a Mongo::DB object.



22
23
24
25
26
27
28
29
# File 'lib/resque.rb', line 22

def mongo=(database)
  if database.is_a?(Mongo::DB)
    @mongo = database
    initialize_mongo
  else
    raise ArgumentError, "Resque.mongo= expects a Mongo::DB database, not a #{database.class}."
  end
end

#mongo_failuresObject



54
55
56
# File 'lib/resque.rb', line 54

def mongo_failures
  mongo['resque.failures']
end

#mongo_statsObject



50
51
52
# File 'lib/resque.rb', line 50

def mongo_stats
  mongo['resque.metrics']
end

#mongo_workersObject



46
47
48
# File 'lib/resque.rb', line 46

def mongo_workers
  mongo['resque.workers']
end

#peek(queue, start = 0, count = 1, mode = :ready) ⇒ Object

Returns an array of items currently queued. Queue name should be a string.

start and count should be integer and can be used for pagination. start is the item to begin, count is how many items to return.

To get the 3rd page of a 30 item, paginatied list one would use:

Resque.peek('my_list', 59, 30)


212
213
214
# File 'lib/resque.rb', line 212

def peek(queue, start = 0, count = 1, mode = :ready)
  list_range(queue, start, count, mode)
end

#pop(queue) ⇒ Object

Pops a job off a queue. Queue name should be a string.

Returns a Ruby object.



167
168
169
170
171
172
173
174
175
176
# File 'lib/resque.rb', line 167

def pop(queue)
  queue = namespace_queue(queue)
  query = {}
  query['delay_until'] = { '$lt' => Time.now } if delayed_queue?(queue)
  #sorting will result in significant performance penalties for large queues, you have been warned.
  item = mongo[queue].find_and_modify(:query => query, :remove => true, :sort => [[:_id, :asc]])
rescue Mongo::OperationFailure => e
  return nil if e.message =~ /No matching object/
  raise e
end

#push(queue, item) ⇒ Object

Pushes a job onto a queue. Queue name should be a string and the item should be any JSON-able Ruby object.

Resque works generally expect the ‘item` to be a hash with the following keys:

class - The String name of the job to run.
 args - An Array of arguments to pass the job. Usually passed
        via `class.to_class.perform(*args)`.

Example

Resque.push('archive', :class => 'Archive', :args => [ 35, 'tar' ])

Returns nothing



158
159
160
161
162
# File 'lib/resque.rb', line 158

def push(queue, item)
  queue = namespace_queue(queue)
  item[:resque_enqueue_timestamp] = Time.now
  mongo[queue] << item
end

#queue_from_class(klass) ⇒ Object

Given a class, try to extrapolate an appropriate queue based on a class instance variable or ‘queue` method.



354
355
356
357
# File 'lib/resque.rb', line 354

def queue_from_class(klass)
  klass.instance_variable_get(:@queue) ||
    (klass.respond_to?(:queue) and klass.queue)
end

#queuesObject

Returns an array of all known Resque queues as strings.



240
241
242
243
244
# File 'lib/resque.rb', line 240

def queues        
  mongo.collection_names.
    select { |name| name =~ /resque\.queues\./ }.
    collect { |name| name.split(".")[2..-1].join('.') }
end

#ready_size(queue) ⇒ Object



194
195
196
197
198
199
200
201
# File 'lib/resque.rb', line 194

def ready_size(queue)
  queue = namespace_queue(queue)
  if delayed_queue? queue
    mongo[queue].find({'delay_until' => { '$lt' => Time.now }}).count
  else
    mongo[queue].count
  end
end

#remove_queue(queue) ⇒ Object

Given a queue name, completely deletes the queue.



253
254
255
256
# File 'lib/resque.rb', line 253

def remove_queue(queue)
  queue = namespace_queue(queue)
  mongo[queue].drop
end

#remove_worker(worker_id) ⇒ Object

A shortcut to unregister_worker useful for command line tool



402
403
404
405
# File 'lib/resque.rb', line 402

def remove_worker(worker_id)
  worker = Resque::Worker.find(worker_id)
  worker.unregister_worker
end

#reserve(queue) ⇒ Object

This method will return a ‘Resque::Job` object or a non-true value depending on whether a job can be obtained. You should pass it the precise name of a queue: case matters.

This method is considered part of the ‘stable` API.



364
365
366
# File 'lib/resque.rb', line 364

def reserve(queue)
  Job.reserve(queue)
end

#size(queue) ⇒ Object

Returns an integer representing the size of a queue. Queue name should be a string.



180
181
182
183
# File 'lib/resque.rb', line 180

def size(queue)
  queue = namespace_queue(queue)
  mongo[queue].count
end

#to_sObject



105
106
107
108
# File 'lib/resque.rb', line 105

def to_s
  connection_info = mongo.connection.primary_pool
  "Resque Client connected to #{connection_info.host}:#{connection_info.port}/#{mongo.name}"
end

#validate(klass, queue = nil) ⇒ Object

Validates if the given klass could be a valid Resque job

If no queue can be inferred this method will raise a ‘Resque::NoQueueError`

If given klass is nil this method will raise a ‘Resque::NoClassError`



373
374
375
376
377
378
379
380
381
382
383
# File 'lib/resque.rb', line 373

def validate(klass, queue = nil)
  queue ||= queue_from_class(klass)

  if !queue
    raise NoQueueError.new("Jobs must be placed onto a queue.")
  end

  if klass.to_s.empty?
    raise NoClassError.new("Jobs must be given a class.")
  end
end

#workersObject

A shortcut to Worker.all



391
392
393
# File 'lib/resque.rb', line 391

def workers
  Worker.all
end

#workingObject

A shortcut to Worker.working



396
397
398
# File 'lib/resque.rb', line 396

def working
  Worker.working
end