Class: Delayed::Backend::ActiveRecord::Job

Inherits:
AbstractJob show all
Includes:
Base
Defined in:
lib/delayed/backend/active_record.rb

Overview

A job object that is persisted to the database. Contains the work object as a YAML field.

Direct Known Subclasses

Failed

Defined Under Namespace

Classes: Failed

Constant Summary

Constants included from Base

Base::ON_HOLD_BLOCKER, Base::ON_HOLD_COUNT, Base::ON_HOLD_LOCKED_BY

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Methods included from Base

#batch?, #expired?, #failed?, #full_name, #hold!, included, #inferred_max_attempts, #initialize_defaults, #invoke_job, #locked?, #name, #on_hold?, #payload_object, #payload_object=, #permanent_failure, #reschedule, #reschedule_at, #unhold!, #unlock

Methods inherited from ActiveRecord::Base

#encode_with, load_for_delayed_job

Instance Attribute Details

#enqueue_result ⇒ Object

Returns the value of attribute enqueue_result.



28
29
30
 
# File 'lib/delayed/backend/active_record.rb', line 28

def enqueue_result
  @enqueue_result
end

Class Method Details

.advisory_lock(lock_name) ⇒ Object 



53
54
55
56
 
# File 'lib/delayed/backend/active_record.rb', line 53

def advisory_lock(lock_name)
  fn_name = connection.quote_table_name("half_md5_as_bigint")
  connection.execute("SELECT pg_advisory_xact_lock(#{fn_name}('#{lock_name}'));")
end

.all_available(queue = Delayed::Settings.queue, min_priority = nil, max_priority = nil, forced_latency: nil) ⇒ Object 



461
462
463
464
465
466
467
468
469
470
471
472
473
474
 
# File 'lib/delayed/backend/active_record.rb', line 461

def self.all_available(queue = Delayed::Settings.queue,
                       min_priority = nil,
                       max_priority = nil,
                       forced_latency: nil)
  min_priority ||= Delayed::MIN_PRIORITY
  max_priority ||= Delayed::MAX_PRIORITY

  check_queue(queue)
  check_priorities(min_priority, max_priority)

  ready_to_run(forced_latency: forced_latency)
    .where(priority: min_priority..max_priority, queue: queue)
    .by_priority
end

.apply_temp_strand!(job_scope, max_concurrent: 1) ⇒ Object

given a scope of non-stranded queued jobs, apply a temporary strand to throttle their execution returns [job_count, new_strand] (this is designed for use in a Rails console or the Canvas Jobs interface)



339
340
341
342
343
344
345
346
347
348
349
350
351
352
 
# File 'lib/delayed/backend/active_record.rb', line 339

def self.apply_temp_strand!(job_scope, max_concurrent: 1)
  if job_scope.where("strand IS NOT NULL OR singleton IS NOT NULL").exists?
    raise ArgumentError, "can't apply strand to already stranded jobs"
  end

  job_count = 0
  new_strand = "tmp_strand_#{SecureRandom.alphanumeric(16)}"
  ::Delayed::Job.transaction do
    job_count = job_scope.update_all(strand: new_strand, max_concurrent: max_concurrent, next_in_strand: false)
    ::Delayed::Job.where(strand: new_strand).order(:id).limit(max_concurrent).update_all(next_in_strand: true)
  end

  [job_count, new_strand]
end

.attempt_advisory_lock(lock_name) ⇒ Object 



48
49
50
51
 
# File 'lib/delayed/backend/active_record.rb', line 48

def attempt_advisory_lock(lock_name)
  fn_name = connection.quote_table_name("half_md5_as_bigint")
  connection.select_value("SELECT pg_try_advisory_xact_lock(#{fn_name}('#{lock_name}'));")
end

.bulk_update(action, opts) ⇒ Object

perform a bulk update of a set of jobs action is :hold, :unhold, or :destroy to specify the jobs to act on, either pass opts = [list of job ids] or opts = <some flavor> to perform on all jobs of that flavor



285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
 
# File 'lib/delayed/backend/active_record.rb', line 285

def self.bulk_update(action, opts)
  raise("Can't #{action} failed jobs") if opts[:flavor].to_s == "failed" && action.to_s != "destroy"

  scope = if opts[:ids]
            if opts[:flavor] == "failed"
              Delayed::Job::Failed.where(id: opts[:ids])
            else
              where(id: opts[:ids])
            end
          elsif opts[:flavor]

            scope_for_flavor(opts[:flavor], opts[:query])
          end

  return 0 unless scope

  case action.to_s
  when "hold"
    scope = scope.where(locked_by: nil)
    scope.update_all(locked_by: ON_HOLD_LOCKED_BY, locked_at: db_time_now, attempts: ON_HOLD_COUNT)
  when "unhold"
    now = db_time_now
    scope = scope.where(locked_by: ON_HOLD_LOCKED_BY)
    scope.update_all([<<~SQL.squish, now, now])
      locked_by=NULL, locked_at=NULL, attempts=0, run_at=(CASE WHEN run_at > ? THEN run_at ELSE ? END), failed_at=NULL
    SQL
  when "destroy"
    scope = scope.where("locked_by IS NULL OR locked_by=?", ON_HOLD_LOCKED_BY) unless opts[:flavor] == "failed"
    scope.delete_all
  end
end

.by_priority ⇒ Object 



216
217
218
 
# File 'lib/delayed/backend/active_record.rb', line 216

def self.by_priority
  order(:priority, :run_at, :id)
end

.clear_locks!(worker_name) ⇒ Object

When a worker is exiting, make sure we don’t have any locked jobs.



221
222
223
 
# File 'lib/delayed/backend/active_record.rb', line 221

def self.clear_locks!(worker_name)
  where(locked_by: worker_name).update_all(locked_by: nil, locked_at: nil)
end

.create(attributes, &block) ⇒ Object 



41
42
43
44
45
46
 
# File 'lib/delayed/backend/active_record.rb', line 41

def create(attributes, &block)
  on_conflict = attributes.delete(:on_conflict)
  # modified from ActiveRecord::Persistence.create and ActiveRecord::Persistence#_insert_record
  job = new(attributes, &block)
  job.single_step_create(on_conflict: on_conflict)
end

.create_singleton(options) ⇒ Object

Create the job on the specified strand, but only if there aren’t any other non-running jobs on that strand. (in other words, the job will still be created if there’s another job on the strand but it’s already running)



480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
 
# File 'lib/delayed/backend/active_record.rb', line 480

def self.create_singleton(options)
  strand = options[:singleton]
  on_conflict = options.delete(:on_conflict) || :use_earliest

  transaction_for_singleton(singleton, on_conflict) do
    job = where(strand: strand, locked_at: nil).next_in_strand_order.first
    new_job = new(options)
    if job
      new_job.initialize_defaults
      job.run_at =
        case on_conflict
        when :use_earliest, :patient
          [job.run_at, new_job.run_at].min
        when :overwrite
          new_job.run_at
        when :loose
          job.run_at
        end
      job.handler = new_job.handler if on_conflict == :overwrite
      job.save! if job.changed?
    else
      new_job.save!
    end
    job || new_job
  end
end

.current ⇒ Object 



187
188
189
 
# File 'lib/delayed/backend/active_record.rb', line 187

def self.current
  where("run_at<=?", db_time_now)
end

.failed ⇒ Object 



195
196
197
 
# File 'lib/delayed/backend/active_record.rb', line 195

def self.failed
  where.not(failed_at: nil)
end

.find_available(limit, queue = Delayed::Settings.queue, min_priority = nil, max_priority = nil) ⇒ Object 



454
455
456
457
458
459
 
# File 'lib/delayed/backend/active_record.rb', line 454

def self.find_available(limit,
                        queue = Delayed::Settings.queue,
                        min_priority = nil,
                        max_priority = nil)
  all_available(queue, min_priority, max_priority).limit(limit).to_a
end

.future ⇒ Object 



191
192
193
 
# File 'lib/delayed/backend/active_record.rb', line 191

def self.future
  where("run_at>?", db_time_now)
end

.get_and_lock_next_available(worker_names, queue = Delayed::Settings.queue, min_priority = nil, max_priority = nil, prefetch: 0, prefetch_owner: nil, forced_latency: nil) ⇒ Object 



362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
 
# File 'lib/delayed/backend/active_record.rb', line 362

def self.get_and_lock_next_available(worker_names,
                                     queue = Delayed::Settings.queue,
                                     min_priority = nil,
                                     max_priority = nil,
                                     prefetch: 0,
                                     prefetch_owner: nil,
                                     forced_latency: nil)

  check_queue(queue)
  check_priorities(min_priority, max_priority)

  loop do
    jobs = maybe_silence_periodic_log do
      if connection.adapter_name == "PostgreSQL" && !Settings.select_random_from_batch
        # In Postgres, we can lock a job and return which row was locked in a single
        # query by using RETURNING. Combine that with the ROW_NUMBER() window function
        # to assign a distinct locked_at value to each job locked, when doing multiple
        # jobs in a single query.
        effective_worker_names = Array(worker_names)

        lock = nil
        lock = "FOR UPDATE SKIP LOCKED" if connection.postgresql_version >= 90_500
        target_jobs = all_available(queue,
                                    min_priority,
                                    max_priority,
                                    forced_latency: forced_latency)
                      .limit(effective_worker_names.length + prefetch)
                      .lock(lock)
        jobs_with_row_number = all.from(target_jobs)
                                  .select("id, ROW_NUMBER() OVER () AS row_number")
        updates = +"locked_by = CASE row_number "
        effective_worker_names.each_with_index do |worker, i|
          updates << "WHEN #{i + 1} THEN #{connection.quote(worker)} "
        end
        updates << "ELSE #{connection.quote(prefetch_owner)} " if prefetch_owner
        updates << "END, locked_at = #{connection.quote(db_time_now)}"

        # Originally this was done with a subquery, but this allows the query planner to
        # side-step the LIMIT. We use a CTE here to force the subquery to be materialized
        # before running the UPDATE.
        #
        # For more details, see:
        #  * https://dba.stackexchange.com/a/69497/55285
        #  * https://github.com/feikesteenbergen/demos/blob/b7ecee8b2a79bf04cbcd74972e6bfb81903aee5d/bugs/update_limit_bug.txt
        query = <<~SQL.squish
          WITH limited_jobs AS (#{jobs_with_row_number.to_sql})
          UPDATE #{quoted_table_name} SET #{updates} FROM limited_jobs WHERE limited_jobs.id=#{quoted_table_name}.id
          RETURNING #{quoted_table_name}.*
        SQL

        jobs = find_by_sql(query)
        # because this is an atomic query, we don't have to return more jobs than we needed
        # to try and lock them, nor is there a possibility we need to try again because
        # all of the jobs we tried to lock had already been locked by someone else
        return jobs.first unless worker_names.is_a?(Array)

        result = jobs.index_by(&:locked_by)
        # all of the prefetched jobs can come back as an array
        result[prefetch_owner] = jobs.select { |j| j.locked_by == prefetch_owner } if prefetch_owner
        return result
      else
        batch_size = Settings.fetch_batch_size
        batch_size *= worker_names.length if worker_names.is_a?(Array)
        find_available(batch_size, queue, min_priority, max_priority)
      end
    end
    if jobs.empty?
      return worker_names.is_a?(Array) ? {} : nil
    end

    jobs = jobs.sort_by { rand } if Settings.select_random_from_batch
    if worker_names.is_a?(Array)
      result = {}
      jobs.each do |job|
        break if worker_names.empty?

        worker_name = worker_names.first
        if job.send(:lock_exclusively!, worker_name)
          result[worker_name] = job
          worker_names.shift
        end
      end
      return result
    else
      locked_job = jobs.detect do |job|
        job.send(:lock_exclusively!, worker_names)
      end
      return locked_job if locked_job
    end
  end
end

.jobs_count(flavor, query = nil) ⇒ Object

get the total job count for the given flavor see list_jobs for documentation on arguments



275
276
277
278
279
 
# File 'lib/delayed/backend/active_record.rb', line 275

def self.jobs_count(flavor,
                    query = nil)
  scope = scope_for_flavor(flavor, query)
  scope.count
end

.list_jobs(flavor, limit, offset = 0, query = nil) ⇒ Object

get a list of jobs of the given flavor in the given queue flavor is :current, :future, :failed, :strand or :tag depending on the flavor, query has a different meaning: for :current and :future, it’s the queue name (defaults to Delayed::Settings.queue) for :strand it’s the strand name for :tag it’s the tag name for :failed it’s ignored



264
265
266
267
268
269
270
271
 
# File 'lib/delayed/backend/active_record.rb', line 264

def self.list_jobs(flavor,
                   limit,
                   offset = 0,
                   query = nil)
  scope = scope_for_flavor(flavor, query)
  order = (flavor.to_s == "future") ? "run_at" : "id desc"
  scope.order(order).limit(limit).offset(offset).to_a
end

.maybe_silence_periodic_log(&block) ⇒ Object 



354
355
356
357
358
359
360
 
# File 'lib/delayed/backend/active_record.rb', line 354

def self.maybe_silence_periodic_log(&block)
  if Settings.silence_periodic_log
    ::ActiveRecord::Base.logger.silence(&block)
  else
    yield
  end
end

.n_strand_options(strand_name, num_strands) ⇒ Object

This overwrites the previous behavior so rather than changing the strand and balancing at queue time, this keeps the strand intact and uses triggers to limit the number running



183
184
185
 
# File 'lib/delayed/backend/active_record.rb', line 183

def self.n_strand_options(strand_name, num_strands)
  { strand: strand_name, max_concurrent: num_strands }
end

.prefetch_jobs_lock_name ⇒ Object 



512
513
514
 
# File 'lib/delayed/backend/active_record.rb', line 512

def self.prefetch_jobs_lock_name
  "Delayed::Job.unlock_orphaned_prefetched_jobs"
end

.processes_locked_locally(name: nil) ⇒ Object 



507
508
509
510
 
# File 'lib/delayed/backend/active_record.rb', line 507

def self.processes_locked_locally(name: nil)
  name ||= Socket.gethostname rescue x
  where("locked_by LIKE ?", "#{name}:%").pluck(:locked_by).map { |locked_by| locked_by.split(":").last.to_i }
end

.ready_to_run(forced_latency: nil) ⇒ Object

a nice stress test: 10_000.times do |i|

Kernel.delay(strand: 's1', run_at: (24.hours.ago + (rand(24.hours.to_i))).system("echo #{i} >> test1.txt")

end 500.times { |i| “ohai”.delay(run_at: (12.hours.ago + (rand(24.hours.to_i))).reverse } then fire up your workers you can check out strand correctness: diff test1.txt <(sort -n test1.txt)



210
211
212
213
214
 
# File 'lib/delayed/backend/active_record.rb', line 210

def self.ready_to_run(forced_latency: nil)
  now = db_time_now
  now -= forced_latency if forced_latency
  where("run_at<=? AND locked_at IS NULL AND next_in_strand=?", now, true)
end

.reconnect! ⇒ Object 



32
33
34
35
36
37
38
 
# File 'lib/delayed/backend/active_record.rb', line 32

def self.reconnect!
  if Rails.version < "6.1"
    ::ActiveRecord::Base.connection_handler.clear_all_connections!
  else
    ::ActiveRecord::Base.connection_handler.clear_all_connections!(nil)
  end
end

.running ⇒ Object 



199
200
201
 
# File 'lib/delayed/backend/active_record.rb', line 199

def self.running
  where("locked_at IS NOT NULL AND locked_by<>'on hold'")
end

.running_jobs ⇒ Object 



229
230
231
 
# File 'lib/delayed/backend/active_record.rb', line 229

def self.running_jobs
  running.order(:locked_at)
end

.scope_for_flavor(flavor, query) ⇒ Object 



233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
 
# File 'lib/delayed/backend/active_record.rb', line 233

def self.scope_for_flavor(flavor, query)
  scope = case flavor.to_s
          when "current"
            current
          when "future"
            future
          when "failed"
            Delayed::Job::Failed
          when "strand"
            where(strand: query)
          when "tag"
            where(tag: query)
          else
            raise ArgumentError, "invalid flavor: #{flavor.inspect}"
          end

  if %w[current future].include?(flavor.to_s)
    queue = query.presence || Delayed::Settings.queue
    scope = scope.where(queue: queue)
  end

  scope
end

.strand_size(strand) ⇒ Object 



225
226
227
 
# File 'lib/delayed/backend/active_record.rb', line 225

def self.strand_size(strand)
  where(strand: strand).count
end

.tag_counts(flavor, limit, offset = 0) ⇒ Object

returns a list of hashes { :tag => tag_name, :count => current_count } in descending count order flavor is :current or :all

Raises:

  • (ArgumentError) 


320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
 
# File 'lib/delayed/backend/active_record.rb', line 320

def self.tag_counts(flavor,
                    limit,
                    offset = 0)
  raise(ArgumentError, "invalid flavor: #{flavor}") unless %w[current all].include?(flavor.to_s)

  scope = case flavor.to_s
          when "current"
            current
          when "all"
            self
          end

  scope = scope.group(:tag).offset(offset).limit(limit)
  scope.order(Arel.sql("COUNT(tag) DESC")).count.map { |t, c| { tag: t, count: c } }
end

.unlock(jobs) ⇒ Object 



527
528
529
530
531
 
# File 'lib/delayed/backend/active_record.rb', line 527

def self.unlock(jobs)
  unlocked = where(id: jobs).update_all(locked_at: nil, locked_by: nil)
  jobs.each(&:unlock)
  unlocked
end

.unlock_orphaned_prefetched_jobs ⇒ Object 



516
517
518
519
520
521
522
523
524
525
 
# File 'lib/delayed/backend/active_record.rb', line 516

def self.unlock_orphaned_prefetched_jobs
  transaction do
    # for db performance reasons, we only need one process doing this at a time
    # so if we can't get an advisory lock, just abort. we'll try again soon
    next unless attempt_advisory_lock(prefetch_jobs_lock_name)

    horizon = db_time_now - (Settings.parent_process[:prefetched_jobs_timeout] * 4)
    where("locked_by LIKE 'prefetch:%' AND locked_at<?", horizon).update_all(locked_at: nil, locked_by: nil)
  end
end

Instance Method Details

#create_and_lock!(worker) ⇒ Object 



579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
 
# File 'lib/delayed/backend/active_record.rb', line 579

def create_and_lock!(worker)
  raise "job already exists" unless new_record?

  # we don't want to process unique constraint violations of
  # running singleton jobs; always do it as two steps
  if singleton
    single_step_create
    lock_exclusively!(worker)
    return
  end

  self.locked_at = Delayed::Job.db_time_now
  self.locked_by = worker
  single_step_create
end

#destroy ⇒ Object 



151
152
153
154
 
# File 'lib/delayed/backend/active_record.rb', line 151

def destroy
  # skip transaction and callbacks
  destroy_row
end

#fail! ⇒ Object 



595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
 
# File 'lib/delayed/backend/active_record.rb', line 595

def fail!
  attrs = attributes
  attrs["original_job_id"] = attrs.delete("id") if Failed.columns_hash.key?("original_job_id")
  attrs["failed_at"] ||= self.class.db_time_now
  attrs.delete("next_in_strand")
  attrs.delete("max_concurrent")
  self.class.transaction do
    failed_job = Failed.create(attrs)
    destroy
    failed_job
  end
rescue
  # we got an error while failing the job -- we need to at least get
  # the job out of the queue
  destroy
  # re-raise so the worker logs the error, at least
  raise
end

#lock_strand_on_create ⇒ Object 



172
173
174
175
176
177
178
 
# File 'lib/delayed/backend/active_record.rb', line 172

def lock_strand_on_create
  return unless strand.present? && instance_of?(Job)

  fn_name = self.class.connection.quote_table_name("half_md5_as_bigint")
  quoted_strand_name = self.class.connection.quote(strand)
  self.class.connection.execute("SELECT pg_advisory_xact_lock(#{fn_name}(#{quoted_strand_name}))")
end

#single_step_create(on_conflict: nil) ⇒ Object 



59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
 
# File 'lib/delayed/backend/active_record.rb', line 59

def single_step_create(on_conflict: nil)
  connection = self.class.connection

  # a before_save callback that we're skipping
  initialize_defaults

  current_time = current_time_from_proper_timezone

  all_timestamp_attributes_in_model.each do |column|
    _write_attribute(column, current_time) unless attribute_present?(column)
  end

  attribute_names = if Rails.version < "7.0"
                      attribute_names_for_partial_writes
                    else
                      attribute_names_for_partial_inserts
                    end
  attribute_names = attributes_for_create(attribute_names)
  values = attributes_with_values(attribute_names)

  im = if Rails.version < "7.0"
         self.class.arel_table.compile_insert(self.class.send(:_substitute_values, values))
       else
         im = Arel::InsertManager.new(self.class.arel_table)
         im.insert(values.transform_keys { |name| self.class.arel_table[name] })
         im
       end

  lock_and_insert = values["strand"] && instance_of?(Job)
  # can't use prepared statements if we're combining multiple statemenets
  sql, binds = if lock_and_insert
                 connection.unprepared_statement do
                   connection.send(:to_sql_and_binds, im)
                 end
               else
                 connection.send(:to_sql_and_binds, im)
               end
  sql = +sql

  if singleton && instance_of?(Job)
    sql << " ON CONFLICT (singleton) WHERE singleton IS NOT NULL AND locked_by IS NULL DO "
    sql << case on_conflict
           when :patient, :loose
             "NOTHING"
           when :overwrite
             "UPDATE SET run_at=EXCLUDED.run_at, handler=EXCLUDED.handler"
           else # :use_earliest
             "UPDATE SET run_at=EXCLUDED.run_at WHERE EXCLUDED.run_at<delayed_jobs.run_at"
           end
  end

  # https://www.postgresql.org/docs/9.5/libpq-exec.html
  # https://stackoverflow.com/questions/39058213/differentiate-inserted-and-updated-rows-in-upsert-using-system-columns
  sql << " RETURNING id, (xmax = 0) AS inserted"

  if lock_and_insert
    # > Multiple queries sent in a single PQexec call are processed in a single transaction,
    # unless there are explicit BEGIN/COMMIT commands included in the query string to divide
    # it into multiple transactions.
    # but we don't need to lock when inserting into Delayed::Failed
    if values["strand"] && instance_of?(Job)
      fn_name = connection.quote_table_name("half_md5_as_bigint")
      quoted_strand = connection.quote((Rails.version < "7.0") ? values["strand"] : values["strand"].value)
      sql = "SELECT pg_advisory_xact_lock(#{fn_name}(#{quoted_strand})); #{sql}"
    end
    result = connection.execute(sql, "#{self.class} Create")
    self.id = result.values.first&.first
    inserted = result.values.first&.second
    result.clear
  else
    result = connection.exec_query(sql, "#{self.class} Create", binds)
    self.id = connection.send(:last_inserted_id, result)
    inserted = result.rows.first&.second
  end

  self.enqueue_result = if id.present? && inserted
                          :inserted
                        elsif id.present? && !inserted
                          :updated
                        else
                          :dropped
                        end

  # it might not get set if there was an existing record, and we didn't update it
  if id
    @new_record = false
    changes_applied
  end

  self
end

#transfer_lock!(from:, to:) ⇒ Object 



553
554
555
556
557
558
559
560
561
562
563
 
# File 'lib/delayed/backend/active_record.rb', line 553

def transfer_lock!(from:, to:)
  now = self.class.db_time_now
  # We don't own this job so we will update the locked_by name and the locked_at
  affected_rows = self.class.where(id: self, locked_by: from).update_all(locked_at: now, locked_by: to)
  if affected_rows == 1
    mark_as_locked!(now, to)
    true
  else
    false
  end
end