workhorse 2.0.0.rc0 → 2.0.0.rc1

data/test/workhorse/worker_test.rb

@@ -97,14 +97,11 @@ class Workhorse::WorkerTest < WorkhorseTest
     with_worker(pool_size: 5, polling_interval: 0.2) do |w|
       assert_equal 5, w.idle
 
-      sleep 0.05
-      Workhorse.enqueue BasicJob.new(sleep_time: 0.2)
+      # Long enough to be seen running at the 0.1s the retries look.
+      Workhorse.enqueue BasicJob.new(sleep_time: 1)
 
-      sleep 0.25
-      assert_equal 4, w.idle
-
-      sleep 0.2
-      assert_equal 5, w.idle
+      with_retries { assert_equal 4, w.idle }
+      with_retries { assert_equal 5, w.idle }
     end
   end
 
@@ -122,42 +119,37 @@ class Workhorse::WorkerTest < WorkhorseTest
   end
 
   def test_perform
-    with_worker(polling_interval: 0.2) do
-      sleep 0.1
-      Workhorse.enqueue BasicJob.new(sleep_time: 0.1)
-      assert_equal 'waiting', Workhorse::DbJob.first.state
+    # Enqueued before the worker starts, as a running one may pick the job up
+    # before its state can be looked at.
+    Workhorse.enqueue BasicJob.new(sleep_time: 0.1)
+    assert_equal 'waiting', Workhorse::DbJob.first.state
 
-      sleep 0.3
+    work_until(pool_size: 5, polling_interval: 0.2) do
+      assert_equal 'succeeded', Workhorse::DbJob.first.state
     end
-
-    assert_equal 'succeeded', Workhorse::DbJob.first.state
   end
 
   def test_params
     BasicJob.results.clear
 
     Workhorse.enqueue BasicJob.new(some_param: 5, sleep_time: 0)
-    work 0.5
-
-    assert_equal 'succeeded', Workhorse::DbJob.first.state
+    work_until(pool_size: 5, polling_interval: 0.2) { assert_equal 'succeeded', Workhorse::DbJob.first.state }
 
     assert_equal 1, BasicJob.results.count
     assert_equal 5, BasicJob.results.first
   end
 
   def test_term
-    with_worker(polling_interval: 0.2) do |w|
+    with_worker(polling_interval: 0.2, auto_terminate: true) do |w|
       Process.kill 'TERM', Process.pid
-      sleep 0.2
-      w.assert_state! :shutdown
+      with_retries { w.assert_state! :shutdown }
     end
   end
 
   def test_int
-    with_worker(polling_interval: 0.2) do |w|
+    with_worker(polling_interval: 0.2, auto_terminate: true) do |w|
       Process.kill 'INT', Process.pid
-      sleep 0.2
-      w.assert_state! :shutdown
+      with_retries { w.assert_state! :shutdown }
     end
   end
 
@@ -174,6 +166,36 @@ class Workhorse::WorkerTest < WorkhorseTest
     FileUtils.rm_f Workhorse::Worker.shutdown_file_for(Process.pid)
   end
 
+  # A fresh checkout or deployment has no tmp/pids. The soft restart touches
+  # its shutdown file there after it has stopped accepting jobs, so failing on
+  # it used to leave a worker that took no jobs and never exited.
+  def test_soft_restart_without_a_pids_directory
+    FileUtils.rm_rf(Rails.root.join('tmp', 'pids'))
+
+    with_worker(pool_size: 1, polling_interval: 0.2) do |w|
+      Process.kill 'USR1', Process.pid
+
+      with_retries { assert_equal :shutdown, w.state }
+      assert File.exist?(Workhorse::Worker.shutdown_file_for(Process.pid))
+    end
+  ensure
+    FileUtils.mkdir_p(Rails.root.join('tmp', 'pids'))
+    FileUtils.rm_f Workhorse::Worker.shutdown_file_for(Process.pid)
+  end
+
+  def test_heartbeat_without_a_pids_directory
+    FileUtils.rm_rf(Rails.root.join('tmp', 'pids'))
+    ENV['WORKHORSE_DAEMON_WORKER_ID'] = '42'
+
+    Workhorse::Worker.new.heartbeat!
+
+    assert File.exist?(Workhorse::Worker.heartbeat_file_for('42'))
+  ensure
+    ENV.delete('WORKHORSE_DAEMON_WORKER_ID')
+    FileUtils.mkdir_p(Rails.root.join('tmp', 'pids'))
+    FileUtils.rm_f(Workhorse::Worker.heartbeat_file_for('42'))
+  end
+
   def test_soft_restart_when_busy_waits_for_job
     with_worker(pool_size: 1, polling_interval: 0.2) do |w|
       Workhorse.enqueue BasicJob.new(sleep_time: 0.5)
@@ -236,7 +258,7 @@ class Workhorse::WorkerTest < WorkhorseTest
   end
 
   def test_soft_restart_ignored_during_shutdown
-    with_worker(pool_size: 1, polling_interval: 0.2) do |w|
+    with_worker(pool_size: 1, polling_interval: 0.2, auto_terminate: true) do |w|
       Process.kill 'TERM', Process.pid
       with_retries { assert_equal :shutdown, w.state }
 
@@ -252,80 +274,61 @@ class Workhorse::WorkerTest < WorkhorseTest
 
   def test_no_queues
     enqueue_in_multiple_queues
-    work 0.2, polling_interval: 0.2
 
-    jobs = Workhorse::DbJob.order(queue: :asc).to_a
-    assert_equal 'succeeded', jobs[0].state
-    assert_equal 'succeeded', jobs[1].state
-    assert_equal 'succeeded', jobs[2].state
+    work_until(pool_size: 5, polling_interval: 0.2) do
+      assert_equal %w[succeeded succeeded succeeded], jobs_by_queue.map(&:state)
+    end
   end
 
   def test_nil_queue
     enqueue_in_multiple_queues
-    work 0.2, queues: [nil], polling_interval: 0.2
 
-    jobs = Workhorse::DbJob.order(queue: :asc).to_a
-    assert_equal 'succeeded', jobs[0].state
-    assert_equal 'waiting',   jobs[1].state
-    assert_equal 'waiting',   jobs[2].state
+    # A job the worker wrongly took would be taken in the same poll as the
+    # one it rightly took, so is locked by the time that one has succeeded.
+    work_until(queues: [nil], pool_size: 5, polling_interval: 0.2) do
+      assert_equal %w[succeeded waiting waiting], jobs_by_queue.map(&:state)
+    end
   end
 
   def test_queues_with_nil
     enqueue_in_multiple_queues
-    work 0.2, queues: [nil, :q1], polling_interval: 0.2
 
-    jobs = Workhorse::DbJob.order(queue: :asc).to_a
-    assert_equal 'succeeded', jobs[0].state
-    assert_equal 'succeeded', jobs[1].state
-    assert_equal 'waiting',   jobs[2].state
+    work_until(queues: [nil, :q1], pool_size: 5, polling_interval: 0.2) do
+      assert_equal %w[succeeded succeeded waiting], jobs_by_queue.map(&:state)
+    end
   end
 
   def test_queues_without_nil
     enqueue_in_multiple_queues
-    work 0.2, queues: %i[q1 q2], polling_interval: 0.2
 
-    jobs = Workhorse::DbJob.order(queue: :asc).to_a
-    assert_equal 'waiting',   jobs[0].state
-    assert_equal 'succeeded', jobs[1].state
-    assert_equal 'succeeded', jobs[2].state
+    work_until(queues: %i[q1 q2], pool_size: 5, polling_interval: 0.2) do
+      assert_equal %w[waiting succeeded succeeded], jobs_by_queue.map(&:state)
+    end
   end
 
   def test_queue_not_parallel
     Workhorse::DbJob.delete_all
-    Workhorse.enqueue BasicJob.new(sleep_time: 0.2), queue: :q1
-    Workhorse.enqueue BasicJob.new(sleep_time: 0.2), queue: :q1
+    first = Workhorse.enqueue BasicJob.new(sleep_time: 2), queue: :q1
+    second = Workhorse.enqueue BasicJob.new(sleep_time: 2), queue: :q1
 
-    work 0.2, polling_interval: 0.2
-    jobs = Workhorse::DbJob.all.to_a
-    assert_equal 'succeeded', jobs[0].state
-    assert_equal 'waiting',   jobs[1].state
+    assert_queued_one_at_a_time [first], [second]
   end
 
   def test_multiple_queued_same_queue
     # One queue
-    Workhorse.enqueue BasicJob.new(sleep_time: 0.2), queue: :q1
-    Workhorse.enqueue BasicJob.new(sleep_time: 0.2), queue: :q1
+    first = Workhorse.enqueue BasicJob.new(sleep_time: 2), queue: :q1
+    second = Workhorse.enqueue BasicJob.new(sleep_time: 2), queue: :q1
 
-    work 0.2, polling_interval: 0.2
-
-    jobs = Workhorse::DbJob.all.to_a
-    assert_equal 'succeeded', jobs[0].state
-    assert_equal 'waiting',   jobs[1].state
+    assert_queued_one_at_a_time [first], [second]
 
     # Two queues
     Workhorse::DbJob.delete_all
-    Workhorse.enqueue BasicJob.new(sleep_time: 0.2), queue: :q1
-    Workhorse.enqueue BasicJob.new(sleep_time: 0.2), queue: :q1
-    Workhorse.enqueue BasicJob.new(sleep_time: 0.2), queue: :q2
-    Workhorse.enqueue BasicJob.new(sleep_time: 0.2), queue: :q2
-
-    work 0.2, polling_interval: 0.2
+    q1_first = Workhorse.enqueue BasicJob.new(sleep_time: 2), queue: :q1
+    q1_second = Workhorse.enqueue BasicJob.new(sleep_time: 2), queue: :q1
+    q2_first = Workhorse.enqueue BasicJob.new(sleep_time: 2), queue: :q2
+    q2_second = Workhorse.enqueue BasicJob.new(sleep_time: 2), queue: :q2
 
-    jobs = Workhorse::DbJob.order(queue: :asc).to_a
-    assert_equal 'succeeded', jobs[0].state
-    assert_equal 'waiting',   jobs[1].state
-    assert_equal 'succeeded', jobs[2].state
-    assert_equal 'waiting',   jobs[3].state
+    assert_queued_one_at_a_time [q1_first, q2_first], [q1_second, q2_second]
   end
 
   def test_order_with_priorities
@@ -338,7 +341,9 @@ class Workhorse::WorkerTest < WorkhorseTest
     Workhorse.enqueue BasicJob.new(some_param: 1, sleep_time: 0), priority: 0
 
     BasicJob.results.clear
-    work 1, pool_size: 1, polling_interval: 0.1
+    work_until(pool_size: 1, polling_interval: 0.1) do
+      assert_equal 6, BasicJob.results.size
+    end
     assert_equal (1..6).to_a, BasicJob.results
   end
 
@@ -352,17 +357,23 @@ class Workhorse::WorkerTest < WorkhorseTest
   end
 
   def test_perform_at
-    Workhorse.enqueue BasicJob.new(sleep_time: 0), perform_at: Time.now
-    Workhorse.enqueue BasicJob.new(sleep_time: 0), perform_at: Time.now + 600
-    work 0.1, polling_interval: 0.1
+    due = Workhorse.enqueue BasicJob.new(sleep_time: 0), perform_at: Time.now
+    later = Workhorse.enqueue BasicJob.new(sleep_time: 0), perform_at: Time.now + 600
+
+    work_until(polling_interval: 0.1) do
+      assert_equal 'succeeded', due.reload.state
+    end
 
-    jobs = Workhorse::DbJob.all.to_a
-    assert_equal 'succeeded', jobs[0].state
-    assert_equal 'waiting',   jobs[1].state
+    assert_equal 'waiting', later.reload.state
   end
 
   def test_controlled_shutdown
-    Workhorse.max_worker_memory_mb = 100
+    # Relative to what this process already uses, as the daemon's workers are
+    # forks of it: a fixed 100MB is exceeded from the start once the Oracle
+    # client library is loaded, and the worker then stops before its first job.
+    # MemHungryJob allocates 250MB, so it still crosses the line.
+    baseline_mb = `ps -p #{Process.pid} -o rss=`.strip.to_i / 1024
+    Workhorse.max_worker_memory_mb = baseline_mb + 100
     with_daemon do
       pid = with_retries do
         pid = daemon.workers.first.pid
@@ -460,4 +471,26 @@ class Workhorse::WorkerTest < WorkhorseTest
     Workhorse.enqueue BasicJob.new(some_param: :q1), queue: :q1
     Workhorse.enqueue BasicJob.new(some_param: :q2), queue: :q2
   end
+
+  # Jobs ordered by queue, the job without a queue first. Sorted here rather
+  # than with ORDER BY, which puts NULL first on MySQL and last on Oracle.
+  # Within a queue, in the order they were enqueued.
+  def jobs_by_queue
+    return Workhorse::DbJob.all.sort_by { |job| [job.queue.nil? ? 0 : 1, job.queue.to_s, job.id] }
+  end
+
+  # Asserts that while the first job of each queue runs, the ones queued behind
+  # it wait. Looked at while they run rather than after a fixed time: the job
+  # behind is rightly taken as soon as the one ahead has finished.
+  def assert_queued_one_at_a_time(firsts, followers)
+    with_worker(pool_size: 5, polling_interval: 0.2, auto_terminate: false) do
+      with_retries { assert_equal(%w[started] * firsts.size, firsts.map { |job| job.reload.state }) }
+
+      # Room for further polls, each a chance to take a follower wrongly, and
+      # well short of the 2s the jobs ahead take.
+      sleep 0.5
+
+      assert_equal(%w[waiting] * followers.size, followers.map { |job| job.reload.state })
+    end
+  end
 end

data/workhorse.gemspec

@@ -1,14 +1,14 @@
 # -*- encoding: utf-8 -*-
-# stub: workhorse 2.0.0.rc0 ruby lib
+# stub: workhorse 2.0.0.rc1 ruby lib
 
 Gem::Specification.new do |s|
   s.name = "workhorse".freeze
-  s.version = "2.0.0.rc0".freeze
+  s.version = "2.0.0.rc1".freeze
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0".freeze) if s.respond_to? :required_rubygems_version=
   s.require_paths = ["lib".freeze]
   s.authors = ["Sitrox".freeze]
-  s.date = "2026-09-29"
+  s.date = "2026-09-30"
   s.files = [".github/workflows/ruby.yml".freeze, ".gitignore".freeze, ".releaser_config".freeze, ".rubocop.yml".freeze, "CHANGELOG.md".freeze, "FAQ.md".freeze, "Gemfile".freeze, "LICENSE".freeze, "README.md".freeze, "RUBY_VERSION".freeze, "Rakefile".freeze, "VERSION".freeze, "bin/rubocop".freeze, "lib/active_job/queue_adapters/workhorse_adapter.rb".freeze, "lib/generators/workhorse/install_generator.rb".freeze, "lib/generators/workhorse/templates/bin/workhorse.rb".freeze, "lib/generators/workhorse/templates/config/initializers/workhorse.rb".freeze, "lib/generators/workhorse/templates/create_table_jobs.rb".freeze, "lib/generators/workhorse/templates/create_table_workhorse_schedules.rb".freeze, "lib/workhorse.rb".freeze, "lib/workhorse/active_job_extension.rb".freeze, "lib/workhorse/daemon.rb".freeze, "lib/workhorse/daemon/shell_handler.rb".freeze, "lib/workhorse/db_job.rb".freeze, "lib/workhorse/enqueuer.rb".freeze, "lib/workhorse/jobs/cleanup_succeeded_jobs.rb".freeze, "lib/workhorse/jobs/detect_late_schedules_job.rb".freeze, "lib/workhorse/jobs/detect_stale_jobs_job.rb".freeze, "lib/workhorse/jobs/run_active_job.rb".freeze, "lib/workhorse/jobs/run_rails_op.rb".freeze, "lib/workhorse/notifiers/base.rb".freeze, "lib/workhorse/notifiers/file_system.rb".freeze, "lib/workhorse/notifiers/none.rb".freeze, "lib/workhorse/notifiers/redis.rb".freeze, "lib/workhorse/performer.rb".freeze, "lib/workhorse/poller.rb".freeze, "lib/workhorse/pool.rb".freeze, "lib/workhorse/schedule.rb".freeze, "lib/workhorse/schedules.rb".freeze, "lib/workhorse/scoped_env.rb".freeze, "lib/workhorse/worker.rb".freeze, "test/active_job/queue_adapters/workhorse_adapter_test.rb".freeze, "test/lib/db_schema.rb".freeze, "test/lib/jobs.rb".freeze, "test/lib/test_helper.rb".freeze, "test/workhorse/daemon_test.rb".freeze, "test/workhorse/db_job_test.rb".freeze, "test/workhorse/enqueuer_test.rb".freeze, "test/workhorse/notifier_test.rb".freeze, "test/workhorse/performer_test.rb".freeze, "test/workhorse/poller_test.rb".freeze, "test/workhorse/pool_test.rb".freeze, "test/workhorse/schedule_test.rb".freeze, "test/workhorse/worker_test.rb".freeze, "test/workhorse/yjit_test.rb".freeze, "workhorse.gemspec".freeze]
   s.homepage = "https://github.com/sitrox/workhorse".freeze
   s.licenses = ["MIT".freeze]

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: workhorse
 version: !ruby/object:Gem::Version
-  version: 2.0.0.rc0
+  version: 2.0.0.rc1
 platform: ruby
 authors:
 - Sitrox
 bindir: bin
 cert_chain: []
-date: 2026-09-29 00:00:00.000000000 Z
+date: 2026-09-30 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport