Class: Puma::Cluster

Inherits:
Runner show all
Defined in:
lib/puma/cluster.rb

Overview

This class is instantiated by the `Puma::Launcher` and used to boot and serve a Ruby application when puma “workers” are needed i.e. when using multi-processes. For example `$ puma -w 5`

At the core of this class is running an instance of `Puma::Server` which gets created via the `start_server` method from the `Puma::Runner` class that this inherits from.

An instance of this class will spawn the number of processes passed in via the `spawn_workers` method call. Each worker will have it's own instance of a `Puma::Server`.

Defined Under Namespace

Classes: Worker

Instance Method Summary collapse

Methods inherited from Runner

#app, #close_control_listeners, #debug, #development?, #error, #load_and_bind, #log, #output_header, #redirected_io?, #ruby_engine, #start_control, #start_server, #stop_control, #test?

Constructor Details

#initialize(cli, events) ⇒ Cluster

Returns a new instance of Cluster.


22
23
24
25
26
27
28
29
30
# File 'lib/puma/cluster.rb', line 22

def initialize(cli, events)
  super cli, events

  @phase = 0
  @workers = []
  @next_check = Time.now

  @phased_restart = false
end

Instance Method Details

#all_workers_booted?Boolean

Returns:

  • (Boolean)

199
200
201
# File 'lib/puma/cluster.rb', line 199

def all_workers_booted?
  @workers.count { |w| !w.booted? } == 0
end

#check_workersObject


203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
# File 'lib/puma/cluster.rb', line 203

def check_workers
  return if @next_check >= Time.now

  @next_check = Time.now + Const::WORKER_CHECK_INTERVAL

  timeout_workers
  wait_workers
  cull_workers
  spawn_workers

  if all_workers_booted?
    # If we're running at proper capacity, check to see if
    # we need to phase any workers out (which will restart
    # in the right phase).
    #
    w = @workers.find { |x| x.phase != @phase }

    if w
      log "- Stopping #{w.pid} for phased upgrade..."
      unless w.term?
        w.term
        log "- #{w.signal} sent to #{w.pid}..."
      end
    end
  end

  @next_check = [
    @workers.reject(&:term?).map(&:ping_timeout).min,
    @next_check
  ].compact.min
end

#cull_workersObject


177
178
179
180
181
182
183
184
185
186
187
188
189
190
# File 'lib/puma/cluster.rb', line 177

def cull_workers
  diff = @workers.size - @options[:workers]
  return if diff < 1

  debug "Culling #{diff.inspect} workers"

  workers_to_cull = @workers[-diff,diff]
  debug "Workers to cull: #{workers_to_cull.inspect}"

  workers_to_cull.each do |worker|
    log "- Worker #{worker.index} (pid: #{worker.pid}) terminating"
    worker.term
  end
end

#fork_worker!Object


422
423
424
425
426
427
# File 'lib/puma/cluster.rb', line 422

def fork_worker!
  if (worker = @workers.find { |w| w.index == 0 })
    worker.phase += 1
  end
  phased_restart
end

#haltObject


381
382
383
384
# File 'lib/puma/cluster.rb', line 381

def halt
  @status = :halt
  wakeup!
end

#next_worker_indexObject


192
193
194
195
196
197
# File 'lib/puma/cluster.rb', line 192

def next_worker_index
  all_positions =  0...@options[:workers]
  occupied_positions = @workers.map { |w| w.index }
  available_positions = all_positions.to_a - occupied_positions
  available_positions.first
end

#phased_restartObject


360
361
362
363
364
365
366
367
# File 'lib/puma/cluster.rb', line 360

def phased_restart
  return false if @options[:preload_app]

  @phased_restart = true
  wakeup!

  true
end

#preload?Boolean

Returns:

  • (Boolean)

418
419
420
# File 'lib/puma/cluster.rb', line 418

def preload?
  @options[:preload_app]
end

#redirect_ioObject


58
59
60
61
62
# File 'lib/puma/cluster.rb', line 58

def redirect_io
  super

  @workers.each { |x| x.hup }
end

#reload_worker_directoryObject


386
387
388
389
390
# File 'lib/puma/cluster.rb', line 386

def reload_worker_directory
  dir = @launcher.restart_dir
  log "+ Changing to #{dir}"
  Dir.chdir dir
end

#restartObject


355
356
357
358
# File 'lib/puma/cluster.rb', line 355

def restart
  @restart = true
  stop
end

#runObject


482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
# File 'lib/puma/cluster.rb', line 482

def run
  @status = :run

  output_header "cluster"

  log "* Process workers: #{@options[:workers]}"

  before = Thread.list

  if preload?
    log "* Preloading application"
    load_and_bind

    after = Thread.list

    if after.size > before.size
      threads = (after - before)
      if threads.first.respond_to? :backtrace
        log "! WARNING: Detected #{after.size-before.size} Thread(s) started in app boot:"
        threads.each do |t|
          log "! #{t.inspect} - #{t.backtrace ? t.backtrace.first : ''}"
        end
      else
        log "! WARNING: Detected #{after.size-before.size} Thread(s) started in app boot"
      end
    end
  else
    log "* Phased restart available"

    unless @launcher.config.app_configured?
      error "No application configured, nothing to run"
      exit 1
    end

    @launcher.binder.parse @options[:binds], self
  end

  read, @wakeup = Puma::Util.pipe

  setup_signals

  # Used by the workers to detect if the master process dies.
  # If select says that @check_pipe is ready, it's because the
  # master has exited and @suicide_pipe has been automatically
  # closed.
  #
  @check_pipe, @suicide_pipe = Puma::Util.pipe

  # Separate pipe used by worker 0 to receive commands to
  # fork new worker processes.
  @fork_pipe, @fork_writer = Puma::Util.pipe

  log "Use Ctrl-C to stop"

  redirect_io

  Plugins.fire_background

  @launcher.write_state

  start_control

  @master_read, @worker_write = read, @wakeup

  @launcher.config.run_hooks :before_fork, nil, @launcher.events
  nakayoshi_gc

  spawn_workers

  Signal.trap "SIGINT" do
    stop
  end

  @launcher.events.fire_on_booted!

  begin
    while @status == :run
      begin
        if @phased_restart
          start_phased_restart
          @phased_restart = false
        end

        check_workers

        res = IO.select([read], nil, nil, [0, @next_check - Time.now].max)

        if res
          req = read.read_nonblock(1)

          @next_check = Time.now if req == "!"
          next if !req || req == "!"

          result = read.gets
          pid = result.to_i

          if req == "b" || req == "f"
            pid, idx = result.split(':').map(&:to_i)
            w = @workers.find {|x| x.index == idx}
            w.pid = pid if w.pid.nil?
          end

          if w = @workers.find { |x| x.pid == pid }
            case req
            when "b"
              w.boot!
              log "- Worker #{w.index} (pid: #{pid}) booted, phase: #{w.phase}"
              @next_check = Time.now
            when "e"
              # external term, see worker method, Signal.trap "SIGTERM"
              w.instance_variable_set :@term, true
            when "t"
              w.term unless w.term?
            when "p"
              w.ping!(result.sub(/^\d+/,'').chomp)
              @launcher.events.fire(:ping!, w)
            end
          else
            log "! Out-of-sync worker list, no #{pid} worker"
          end
        end

      rescue Interrupt
        @status = :stop
      end
    end

    stop_workers unless @status == :halt
  ensure
    @check_pipe.close
    @suicide_pipe.close
    read.close
    @wakeup.close
  end
end

#setup_signalsObject

We do this in a separate method to keep the lambda scope of the signals handlers as small as possible.


431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
# File 'lib/puma/cluster.rb', line 431

def setup_signals
  if @options[:fork_worker]
    Signal.trap "SIGURG" do
      fork_worker!
    end

    # Auto-fork after the specified number of requests.
    if (fork_requests = @options[:fork_worker].to_i) > 0
      @launcher.events.register(:ping!) do |w|
        fork_worker! if w.index == 0 &&
          w.phase == 0 &&
          w.last_status[:requests_count] >= fork_requests
      end
    end
  end

  Signal.trap "SIGCHLD" do
    wakeup!
  end

  Signal.trap "TTIN" do
    @options[:workers] += 1
    wakeup!
  end

  Signal.trap "TTOU" do
    @options[:workers] -= 1 if @options[:workers] >= 2
    wakeup!
  end

  master_pid = Process.pid

  Signal.trap "SIGTERM" do
    # The worker installs their own SIGTERM when booted.
    # Until then, this is run by the worker and the worker
    # should just exit if they get it.
    if Process.pid != master_pid
      log "Early termination of worker"
      exit! 0
    else
      @launcher.close_binder_listeners

      stop_workers
      stop

      raise(SignalException, "SIGTERM") if @options[:raise_exception_on_sigterm]
      exit 0 # Clean exit, workers were stopped
    end
  end
end

#spawn_worker(idx, master) ⇒ Object


163
164
165
166
167
168
169
170
171
172
173
174
175
# File 'lib/puma/cluster.rb', line 163

def spawn_worker(idx, master)
  @launcher.config.run_hooks :before_worker_fork, idx, @launcher.events

  pid = fork { worker(idx, master) }
  if !pid
    log "! Complete inability to spawn new workers detected"
    log "! Seppuku is the only choice."
    exit! 1
  end

  @launcher.config.run_hooks :after_worker_fork, idx, @launcher.events
  pid
end

#spawn_workersObject


133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
# File 'lib/puma/cluster.rb', line 133

def spawn_workers
  diff = @options[:workers] - @workers.size
  return if diff < 1

  master = Process.pid
  if @options[:fork_worker]
    @fork_writer << "-1\n"
  end

  diff.times do
    idx = next_worker_index

    if @options[:fork_worker] && idx != 0
      @fork_writer << "#{idx}\n"
      pid = nil
    else
      pid = spawn_worker(idx, master)
    end

    debug "Spawned worker: #{pid}"
    @workers << Worker.new(idx, pid, @phase, @options)
  end

  if @options[:fork_worker] &&
    @workers.all? {|x| x.phase == @phase}

    @fork_writer << "0\n"
  end
end

#start_phased_restartObject


47
48
49
50
51
52
53
54
55
56
# File 'lib/puma/cluster.rb', line 47

def start_phased_restart
  @phase += 1
  log "- Starting phased worker restart, phase: #{@phase}"

  # Be sure to change the directory again before loading
  # the app. This way we can pick up new code.
  dir = @launcher.restart_dir
  log "+ Changing to #{dir}"
  Dir.chdir dir
end

#statsObject

Inside of a child process, this will return all zeroes, as @workers is only populated in the master process.


394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
# File 'lib/puma/cluster.rb', line 394

def stats
  old_worker_count = @workers.count { |w| w.phase != @phase }
  worker_status = @workers.map do |w|
    {
      started_at: w.started_at.utc.iso8601,
      pid: w.pid,
      index: w.index,
      phase: w.phase,
      booted: w.booted?,
      last_checkin: w.last_checkin.utc.iso8601,
      last_status: w.last_status,
    }
  end

  {
    started_at: @started_at.utc.iso8601,
    workers: @workers.size,
    phase: @phase,
    booted_workers: worker_status.count { |w| w[:booted] },
    old_workers: old_worker_count,
    worker_status: worker_status,
  }
end

#stopObject


369
370
371
372
# File 'lib/puma/cluster.rb', line 369

def stop
  @status = :stop
  wakeup!
end

#stop_blockedObject


374
375
376
377
378
379
# File 'lib/puma/cluster.rb', line 374

def stop_blocked
  @status = :stop if @status == :run
  wakeup!
  @control.stop(true) if @control
  Process.waitall
end

#stop_workersObject


32
33
34
35
36
37
38
39
40
41
42
43
44
45
# File 'lib/puma/cluster.rb', line 32

def stop_workers
  log "- Gracefully shutting down workers..."
  @workers.each { |x| x.term }

  begin
    loop do
      wait_workers
      break if @workers.reject {|w| w.pid.nil?}.empty?
      sleep 0.2
    end
  rescue Interrupt
    log "! Cancelled waiting for workers"
  end
end

#wakeup!Object


235
236
237
238
239
240
241
242
243
# File 'lib/puma/cluster.rb', line 235

def wakeup!
  return unless @wakeup

  begin
    @wakeup.write "!" unless @wakeup.closed?
  rescue SystemCallError, IOError
    Thread.current.purge_interrupt_queue if Thread.current.respond_to? :purge_interrupt_queue
  end
end

#worker(index, master) ⇒ Object


245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
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
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
# File 'lib/puma/cluster.rb', line 245

def worker(index, master)
  title  = "puma: cluster worker #{index}: #{master}"
  title += " [#{@options[:tag]}]" if @options[:tag] && !@options[:tag].empty?
  $0 = title

  Signal.trap "SIGINT", "IGNORE"
  Signal.trap "SIGCHLD", "DEFAULT"

  fork_worker = @options[:fork_worker] && index == 0

  @workers = []
  if !@options[:fork_worker] || fork_worker
    @master_read.close
    @suicide_pipe.close
    @fork_writer.close
  end

  Thread.new do
    Puma.set_thread_name "worker check pipe"
    IO.select [@check_pipe]
    log "! Detected parent died, dying"
    exit! 1
  end

  # If we're not running under a Bundler context, then
  # report the info about the context we will be using
  if !ENV['BUNDLE_GEMFILE']
    if File.exist?("Gemfile")
      log "+ Gemfile in context: #{File.expand_path("Gemfile")}"
    elsif File.exist?("gems.rb")
      log "+ Gemfile in context: #{File.expand_path("gems.rb")}"
    end
  end

  # Invoke any worker boot hooks so they can get
  # things in shape before booting the app.
  @launcher.config.run_hooks :before_worker_boot, index, @launcher.events

  server = @server ||= start_server
  restart_server = Queue.new << true << false

  if fork_worker
    restart_server.clear
    worker_pids = []
    Signal.trap "SIGCHLD" do
      wakeup! if worker_pids.reject! do |p|
        Process.wait(p, Process::WNOHANG) rescue true
      end
    end

    Thread.new do
      Puma.set_thread_name "worker fork pipe"
      while (idx = @fork_pipe.gets)
        idx = idx.to_i
        if idx == -1 # stop server
          if restart_server.length > 0
            restart_server.clear
            server.begin_restart(true)
            @launcher.config.run_hooks :before_refork, nil, @launcher.events
            nakayoshi_gc
          end
        elsif idx == 0 # restart server
          restart_server << true << false
        else # fork worker
          worker_pids << pid = spawn_worker(idx, master)
          @worker_write << "f#{pid}:#{idx}\n" rescue nil
        end
      end
    end
  end

  Signal.trap "SIGTERM" do
    @worker_write << "e#{Process.pid}\n" rescue nil
    server.stop
    restart_server << false
  end

  begin
    @worker_write << "b#{Process.pid}:#{index}\n"
  rescue SystemCallError, IOError
    Thread.current.purge_interrupt_queue if Thread.current.respond_to? :purge_interrupt_queue
    STDERR.puts "Master seems to have exited, exiting."
    return
  end

  Thread.new(@worker_write) do |io|
    Puma.set_thread_name "stat payload"

    while true
      sleep Const::WORKER_CHECK_INTERVAL
      begin
        require 'json'
        io << "p#{Process.pid}#{server.stats.to_json}\n"
      rescue IOError
        Thread.current.purge_interrupt_queue if Thread.current.respond_to? :purge_interrupt_queue
        break
      end
    end
  end

  server.run.join while restart_server.pop

  # Invoke any worker shutdown hooks so they can prevent the worker
  # exiting until any background operations are completed
  @launcher.config.run_hooks :before_worker_shutdown, index, @launcher.events
ensure
  @worker_write << "t#{Process.pid}\n" rescue nil
  @worker_write.close
end