service_skeleton 0.0.0.1.ENOTAG

data/ultravisor/lib/ultravisor.rb

@@ -0,0 +1,216 @@
+# frozen_string_literal: true
+
+require "logger"
+
+require_relative "./ultravisor/child"
+require_relative "./ultravisor/error"
+require_relative "./ultravisor/logging_helpers"
+
+# A super-dooOOOoooper supervisor.
+#
+class Ultravisor
+  include LoggingHelpers
+
+  def initialize(children: [], strategy: :one_for_one, logger: Logger.new("/dev/null"))
+    @queue, @logger = Queue.new, logger
+
+    @strategy = strategy
+    validate_strategy
+
+    @op_m, @op_cv = Mutex.new, ConditionVariable.new
+    @running_thread = nil
+
+    initialize_children(children)
+  end
+
+  def run
+    logger.debug(logloc) { "called" }
+
+    @op_m.synchronize do
+      if @running_thread
+        raise AlreadyRunningError,
+              "This ultravisor is already running"
+      end
+
+      @queue.clear
+      @running_thread = Thread.current
+      Thread.current.name = "Ultravisor"
+    end
+
+    logger.debug(logloc) { "Going to start children #{@children.map(&:first).inspect}" }
+    @children.each { |c| c.last.spawn(@queue) }
+
+    process_events
+
+    @op_m.synchronize do
+      logger.debug(logloc) { "Shutdown time for #{@children.reverse.map(&:first).inspect}" }
+      @children.reverse.each { |c| c.last.shutdown }
+
+      @running_thread = nil
+      @op_cv.broadcast
+    end
+
+    self
+  end
+
+  def shutdown(wait: true, force: false)
+    @op_m.synchronize do
+      return self unless @running_thread
+      if force
+        @children.reverse.each { |c| c.last.shutdown(force: true) }
+        @running_thread.kill
+        @running_thread = nil
+        @op_cv.broadcast
+      else
+        @queue << :shutdown
+        if wait
+          @op_cv.wait(@op_m) while @running_thread
+        end
+      end
+    end
+    self
+  end
+
+  def [](id)
+    @children.assoc(id)&.last
+  end
+
+  def add_child(**args)
+    logger.debug(logloc) { "Adding child #{args[:id].inspect}" }
+    args[:logger] ||= logger
+
+    @op_m.synchronize do
+      c = Ultravisor::Child.new(**args)
+
+      if @children.assoc(c.id)
+        raise DuplicateChildError,
+              "Child with ID #{c.id.inspect} already exists"
+      end
+
+      @children << [c.id, c]
+
+      if @running_thread
+        logger.debug(logloc) { "Auto-starting new child #{args[:id].inspect}" }
+        c.spawn(@queue)
+      end
+    end
+  end
+
+  def remove_child(id)
+    logger.debug(logloc) { "Removing child #{id.inspect}" }
+
+    @op_m.synchronize do
+      c = @children.assoc(id)
+
+      return nil if c.nil?
+
+      @children.delete(c)
+      if @running_thread
+        logger.debug(logloc) { "Shutting down removed child #{id.inspect}" }
+        c.last.shutdown
+      end
+    end
+  end
+
+  private
+
+  def validate_strategy
+    unless %i{one_for_one all_for_one rest_for_one}.include?(@strategy)
+      raise ArgumentError,
+            "Invalid strategy #{@strategy.inspect}"
+    end
+  end
+
+  def initialize_children(children)
+    unless children.is_a?(Array)
+      raise ArgumentError,
+            "children must be an Array"
+    end
+
+    @children = []
+
+    children.each do |cfg|
+      cfg[:logger] ||= logger
+      c = Ultravisor::Child.new(**cfg)
+      if @children.assoc(c.id)
+        raise DuplicateChildError,
+              "Duplicate child ID: #{c.id.inspect}"
+      end
+
+       @children << [c.id, c]
+    end
+  end
+
+  def process_events
+    loop do
+      qe = @queue.pop
+
+      case qe
+      when Ultravisor::Child
+        logger.debug(logloc) { "Received Ultravisor::Child queue entry for #{qe.id}" }
+        @op_m.synchronize { child_exited(qe) }
+      when :shutdown
+        logger.debug(logloc) { "Received :shutdown queue entry" }
+        break
+      else
+        logger.error(logloc) { "Unknown queue entry: #{qe.inspect}" }
+      end
+    end
+  end
+
+  def child_exited(child)
+    if child.termination_exception
+      log_exception(child.termination_exception, "Ultravisor::Child(#{child.id.inspect})") { "Thread terminated by unhandled exception" }
+    end
+
+    if @running_thread.nil?
+      logger.debug(logloc) { "Child termination after shutdown" }
+      # Child termination processed after we've shut down... nope
+      return
+    end
+
+    begin
+      return unless child.restart?
+    rescue Ultravisor::BlownRestartPolicyError
+      # Uh oh...
+      logger.error(logloc) { "Child #{child.id} has exceeded its restart policy.  Shutting down the Ultravisor." }
+      @queue << :shutdown
+      return
+    end
+
+    case @strategy
+    when :all_for_one
+      @children.reverse.each do |id, c|
+        # Don't need to shut down the child that has caused all this mess
+        next if child.id == id
+
+        c.shutdown
+      end
+    when :rest_for_one
+      @children.reverse.each do |id, c|
+        # Don't go past the child that caused the problems
+        break if child.id == id
+
+        c.shutdown
+      end
+    end
+
+    sleep child.restart_delay
+
+    case @strategy
+    when :all_for_one
+      @children.each do |_, c|
+        c.spawn(@queue)
+      end
+    when :rest_for_one
+      s = false
+      @children.each do |id, c|
+        s = true if child.id == id
+
+        c.spawn(@queue) if s
+      end
+    when :one_for_one
+      child.spawn(@queue)
+    end
+  end
+end

data/ultravisor/lib/ultravisor/child.rb

@@ -0,0 +1,481 @@
+# frozen_string_literal: true
+
+require_relative "./logging_helpers"
+
+class Ultravisor
+  class Child
+    include LoggingHelpers
+
+    attr_reader :id
+
+    def initialize(id:,
+                   klass:,
+                   args: [],
+                   method:,
+                   restart: :always,
+                   restart_policy: {
+                      period: 5,
+                      max: 3,
+                      delay: 1,
+                   },
+                   shutdown: {
+                      method: nil,
+                      timeout: 1,
+                   },
+                   logger: Logger.new("/dev/null"),
+                   enable_castcall: false,
+                   access: nil
+                  )
+
+      @logger = logger
+      @id = id
+
+      @klass, @args, @method = klass, args, method
+      validate_kam
+
+      @restart = restart
+      validate_restart
+
+      @restart_policy = restart_policy
+      validate_restart_policy
+
+      @shutdown_spec = shutdown
+      validate_shutdown_spec
+
+      @access = access
+      validate_access
+
+      @enable_castcall = enable_castcall
+
+      @runtime_history = []
+
+      @spawn_m    = Mutex.new
+      @spawn_cv   = ConditionVariable.new
+
+      @shutdown_m = Mutex.new
+    end
+
+    def spawn(term_queue)
+      @spawn_m.synchronize do
+        @value      = nil
+        @exception  = nil
+        @start_time = Time.now
+        @instance   = new_instance
+
+        @spawn_id = sid = rand
+
+        Thread.handle_interrupt(::Exception => :never, ::Numeric => :never) do
+          @thread = Thread.new do
+            Thread.current.name = "Ultravisor::Child(#{@id})"
+            logger.debug(logloc) { "Spawning new instance of #{@id}" }
+
+            begin
+              Thread.handle_interrupt(::Exception => :immediate, ::Numeric => :immediate) do
+                logger.debug(logloc) { "Calling #{@klass}##{@method} to start #{@id} running" }
+                @value = @instance.public_send(@method)
+              end
+            rescue Exception => ex
+              @exception = ex
+            ensure
+              @spawn_m.synchronize do
+                # Even if a thread gets whacked by Thread#kill, ensure blocks
+                # still get run.  This is... wonderful!  And terrifying!
+
+                termination_cleanup(term_queue) if @spawn_id == sid
+              end
+            end
+          end
+        end
+
+        @spawn_cv.broadcast
+      end
+
+      self
+    end
+
+    def shutdown(force: false)
+      @shutdown_m.synchronize do
+        th  = nil
+        sid = nil
+
+        @spawn_m.synchronize do
+          return if @thread.nil? || @thread == Thread.current
+
+          # Take a reference to the running thread, so we don't need to
+          # keep acquiring spawn_m every time we want to do something
+          # with it -- which causes collisions when it comes time to
+          # wait on the terminating thread, which is itself is trying
+          # to acquire the same lock so it can cleanup.
+          th  = @thread
+          sid = @spawn_id
+
+          # Let everyone know we're in shutdown mode
+          @shutting_down = true
+        end
+
+        if @shutdown_spec[:method] && !force
+          begin
+            @instance.public_send(@shutdown_spec[:method])
+          rescue Exception => ex
+            log_exception(ex) { "Unhandled exception when calling #{@shutdown_spec[:method].inspect} on child #{id}" }
+            th.kill
+          end
+        else
+          th.kill
+        end
+
+        unless th.join(@shutdown_spec[:timeout])
+          logger.info(logloc) { "Child instance for #{self.id} did not cleanly shutdown within #{@shutdown_spec[:timeout]} seconds; force-killing the thread" }
+          th.kill
+        end
+
+        # Last chance, bubs
+        unless th.join(0.1)
+          logger.error(logloc) { "Child thread for #{self.id} appears hung; abandoning thread #{th}" }
+
+          # If we get here, then the worker instance has seized up spectacularly,
+          # and the cleanup in the `spawn` ensure hasn't triggered, so we need
+          # to do the cleanup instead.
+          @spawn_m.synchronize do
+            termination_cleanup if @spawn_id == sid
+          end
+        end
+      end
+    end
+
+    def wait
+      @spawn_m.synchronize do
+        @spawn_cv.wait(@spawn_m) while @thread
+      end
+    end
+
+    def termination_exception
+      @spawn_m.synchronize do
+        @spawn_cv.wait(@spawn_m) while @thread
+        @exception
+      end
+    end
+
+    def termination_value
+      @spawn_m.synchronize do
+        @spawn_cv.wait(@spawn_m) while @thread
+        @value
+      end
+    end
+
+    def restart_delay
+      d = begin
+        case @restart_policy[:delay]
+        when Numeric
+          @restart_policy[:delay]
+        when Range
+          @restart_policy[:delay].first + (@restart_policy[:delay].last - @restart_policy[:delay].first) * rand
+        end
+      end
+
+      [0, d].max
+    end
+
+    def restart?
+      if blown_policy?
+        raise BlownRestartPolicyError,
+              "Child #{self.id} has restarted more than #{@restart_policy[:max]} times in #{@restart_policy[:period]} seconds."
+      end
+
+      !!(@restart == :always || (@restart == :on_failure && termination_exception))
+    end
+
+    def unsafe_instance
+      unless @access == :unsafe
+        raise Ultravisor::ThreadSafetyError,
+              "#unsafe_instance called on a child not declared with access: :unsafe"
+      end
+
+      current_instance
+    end
+
+    def cast
+      unless castcall_enabled?
+        raise NoMethodError,
+              "undefined method `cast' for #{self}"
+      end
+
+      CastReceiver.new do |castback|
+        @spawn_m.synchronize do
+          while @instance.nil?
+            #:nocov:
+            @spawn_cv.wait(@spawn_m)
+            #:nocov:
+          end
+
+          unless @instance.respond_to? castback.method_name
+            raise NoMethodError,
+                  "undefined method `#{castback.method_name}' for #{@instance}"
+          end
+
+          begin
+            @instance.instance_variable_get(:@ultravisor_child_castcall_queue) << castback
+          rescue ClosedQueueError
+            # casts aren't guaranteed to ever execute, so dropping it
+            # when the instance's queue has closed is perfectly valid
+          end
+
+          @castcall_fd_writer.putc "!"
+        end
+      end
+    end
+
+    def call
+      unless castcall_enabled?
+        raise NoMethodError,
+              "undefined method `call' for #{self}"
+      end
+
+      CallReceiver.new do |callback|
+        @spawn_m.synchronize do
+          while @instance.nil?
+            #:nocov:
+            @spawn_cv.wait(@spawn_m)
+            #:nocov:
+          end
+
+          unless @instance.respond_to? callback.method_name
+            raise NoMethodError,
+                  "undefined method `#{callback.method_name}' for #{@instance}"
+          end
+
+          begin
+            @instance.instance_variable_get(:@ultravisor_child_castcall_queue) << callback
+          rescue ClosedQueueError
+            raise ChildRestartedError
+          end
+
+          @castcall_fd_writer.putc "!"
+        end
+      end
+    end
+
+    private
+
+    def validate_kam
+      if @klass.instance_method(:initialize).arity == 0 && @args != []
+        raise InvalidKAMError,
+              "#{@klass.to_s}.new takes no arguments, but args not empty."
+      end
+
+      begin
+        if @klass.instance_method(@method).arity != 0
+          raise InvalidKAMError,
+                "#{@klass.to_s}##{@method} must not take arguments"
+        end
+      rescue NameError
+        raise InvalidKAMError,
+              "#{@klass.to_s} has no instance method #{@method}"
+      end
+    end
+
+    def validate_restart
+      unless %i{never on_failure always}.include?(@restart)
+        raise ArgumentError,
+              "Invalid value for restart: #{@restart.inspect}"
+      end
+    end
+
+    def validate_restart_policy
+      unless @restart_policy.is_a?(Hash)
+        raise ArgumentError,
+              "restart_policy must be a hash (got #{@restart_policy.inspect})"
+      end
+
+      bad_keys = @restart_policy.keys - %i{period max delay}
+      unless bad_keys.empty?
+        raise ArgumentError,
+              "Invalid key(s) in restart_policy: #{bad_keys.inspect}"
+      end
+
+      # Restore any missing defaults
+      @restart_policy = { period: 5, max: 3, delay: 1 }.merge(@restart_policy)
+
+      unless @restart_policy[:period].is_a?(Numeric) && @restart_policy[:period].positive?
+        raise ArgumentError,
+              "Invalid restart_policy period #{@restart_policy[:period].inspect}: must be positive integer"
+      end
+
+      unless @restart_policy[:max].is_a?(Numeric) && !@restart_policy[:max].negative?
+        raise ArgumentError,
+              "Invalid restart_policy max #{@restart_policy[:period].inspect}: must be non-negative integer"
+      end
+
+      case @restart_policy[:delay]
+      when Numeric
+        if @restart_policy[:delay].negative?
+          raise ArgumentError,
+                "Invalid restart_policy delay #{@restart_policy[:delay].inspect}: must be non-negative integer or range"
+        end
+      when Range
+        if @restart_policy[:delay].first >= @restart_policy[:delay].last
+          raise ArgumentError,
+                "Invalid restart_policy delay #{@restart_policy[:delay].inspect}: must be non-negative integer or increasing range"
+        end
+
+        if @restart_policy[:delay].first.negative?
+          raise ArgumentError,
+                "Invalid restart_policy delay #{@restart_policy[:delay].inspect}: range must not be negative"
+        end
+      else
+        raise ArgumentError,
+              "Invalid restart_policy delay #{@restart_policy[:delay].inspect}: must be non-negative integer or range"
+      end
+    end
+
+    def validate_shutdown_spec
+      unless @shutdown_spec.is_a?(Hash)
+        raise ArgumentError,
+              "shutdown must be a hash (got #{@shutdown_spec.inspect})"
+      end
+
+      bad_keys = @shutdown_spec.keys - %i{method timeout}
+      unless bad_keys.empty?
+        raise ArgumentError,
+              "Invalid key(s) in shutdown specification: #{bad_keys.inspect}"
+      end
+
+      # Restore any missing defaults
+      @shutdown_spec = { method: nil, timeout: 1 }.merge(@shutdown_spec)
+
+      if @shutdown_spec[:method]
+        begin
+          unless @klass.instance_method(@shutdown_spec[:method]).arity == 0
+            raise ArgumentError,
+                  "Shutdown method #{@klass.to_s}##{@shutdown_spec[:method]} must not take any arguments"
+          end
+        rescue NameError
+          raise ArgumentError,
+                "Shutdown method #{@klass.to_s}##{@shutdown_spec[:method]} is not defined"
+        end
+      end
+
+      unless @shutdown_spec[:timeout].is_a?(Numeric) && !@shutdown_spec[:timeout].negative?
+        raise ArgumentError,
+              "Invalid shutdown timeout #{@shutdown_spec[:timeout].inspect}: must be non-negative integer"
+      end
+    end
+
+    def validate_access
+      return if @access.nil?
+
+      unless %i{unsafe}.include? @access
+        raise ArgumentError,
+              "Invalid instance access specification: #{@access.inspect}"
+      end
+    end
+
+    def castcall_enabled?
+      !!@enable_castcall
+    end
+
+    def new_instance
+      # If there is anything that pisses me off about Ruby's varargs handling more
+      # than the fact that *[] is an empty array, and not a zero-length argument
+      # list, I don't know what it is.  Everything else works *so well*, and this...
+      # urgh.
+      if @klass.instance_method(:initialize).arity == 0
+        @klass.new()
+      else
+        @klass.new(*@args)
+      end.tap do |i|
+        if castcall_enabled?
+          i.singleton_class.prepend(Ultravisor::Child::ProcessCastCall)
+          i.instance_variable_set(:@ultravisor_child_castcall_queue, Queue.new)
+
+          r, @castcall_fd_writer = IO.pipe
+          i.instance_variable_set(:@ultravisor_child_castcall_fd, r)
+        end
+      end
+    end
+
+    def current_instance
+      @spawn_m.synchronize do
+        while @instance.nil?
+          @spawn_cv.wait(@spawn_m)
+        end
+
+        return @instance
+      end
+    end
+
+    def blown_policy?
+      cumulative_runtime = 0
+      # This starts at 1 because we only check this during a restart, so
+      # by definition there must have been at least one recent restart
+      recent_restart_count = 1
+
+      @runtime_history.each do |t|
+        cumulative_runtime += t
+
+        if cumulative_runtime < @restart_policy[:period]
+          recent_restart_count += 1
+        end
+      end
+
+      logger.debug(logloc) { "@runtime_history: #{@runtime_history.inspect}, cumulative_runtime: #{cumulative_runtime}, recent_restart_count: #{recent_restart_count}, restart_policy: #{@restart_policy.inspect}" }
+
+      if recent_restart_count > @restart_policy[:max]
+        return true
+      end
+
+      @runtime_history = @runtime_history[0..recent_restart_count]
+
+      false
+    end
+
+    def termination_cleanup(term_queue = nil)
+      unless @spawn_m.owned?
+        #:nocov:
+        raise ThreadSafetyError,
+              "termination_cleanup must be called while holding the @spawn_m lock"
+        #:nocov:
+      end
+
+      if @start_time
+        @runtime_history.unshift(Time.now.to_f - @start_time.to_f)
+        @start_time = nil
+      end
+
+      term_queue << self if term_queue && !@shutting_down
+
+      if castcall_enabled?
+        cc_q = @instance.instance_variable_get(:@ultravisor_child_castcall_queue)
+        cc_q.close
+        x = 0
+        begin
+          loop do
+            cc_q.pop(true).child_restarted!
+          end
+        rescue ThreadError => ex
+          raise unless ex.message == "queue empty"
+        end
+
+        @instance.instance_variable_get(:@ultravisor_child_castcall_fd).close
+        @instance.instance_variable_set(:@ultravisor_child_castcall_fd, nil)
+        @castcall_fd_writer.close
+        @castcall_fd_writer = nil
+      end
+
+      @instance = nil
+
+      if @thread
+        @thread = nil
+        @spawn_cv.broadcast
+      end
+
+      @spawn_id = nil
+    end
+  end
+end
+
+require_relative "./child/call"
+require_relative "./child/call_receiver"
+require_relative "./child/cast"
+require_relative "./child/cast_receiver"
+require_relative "./child/process_cast_call"