# Copyright 2009 Brian Quinlan. All Rights Reserved. # Licensed to PSF under a Contributor Agreement. """Implements ProcessPoolExecutor. The follow diagram and text describe the data-flow through the system: |======================= In-process =====================|== Out-of-process ==| +----------+ +----------+ +--------+ +-----------+ +---------+ | | => | Work Ids | | | | Call Q | | Process | | | +----------+ | | +-----------+ | Pool | | | | ... | | | | ... | +---------+ | | | 6 | => | | => | 5, call() | => | | | | | 7 | | | | ... | | | | Process | | ... | | Local | +-----------+ | Process | | Pool | +----------+ | Worker | | #1..n | | Executor | | Thread | | | | | +----------- + | | +-----------+ | | | | <=> | Work Items | <=> | | <= | Result Q | <= | | | | +------------+ | | +-----------+ | | | | | 6: call() | | | | ... | | | | | | future | | | | 4, result | | | | | | ... | | | | 3, except | | | +----------+ +------------+ +--------+ +-----------+ +---------+ Executor.submit() called: - creates a uniquely numbered _WorkItem and adds it to the "Work Items" dict - adds the id of the _WorkItem to the "Work Ids" queue Local worker thread: - reads work ids from the "Work Ids" queue and looks up the corresponding WorkItem from the "Work Items" dict: if the work item has been cancelled then it is simply removed from the dict, otherwise it is repackaged as a _CallItem and put in the "Call Q". New _CallItems are put in the "Call Q" until "Call Q" is full. NOTE: the size of the "Call Q" is kept small because calls placed in the "Call Q" can no longer be cancelled with Future.cancel(). - reads _ResultItems from "Result Q", updates the future stored in the "Work Items" dict and deletes the dict entry Process #1..n: - reads _CallItems from "Call Q", executes the calls, and puts the resulting _ResultItems in "Result Q" """ __author__ = 'Brian Quinlan (brian@sweetapp.com)' import atexit import os from concurrent.futures import _base import queue from queue import Full import multiprocessing as mp from multiprocessing.connection import wait from multiprocessing.queues import Queue import threading import weakref from functools import partial import itertools import sys import traceback # Workers are created as daemon threads and processes. This is done to allow the # interpreter to exit when there are still idle processes in a # ProcessPoolExecutor's process pool (i.e. shutdown() was not called). However, # allowing workers to die with the interpreter has two undesirable properties: # - The workers would still be running during interpreter shutdown, # meaning that they would fail in unpredictable ways. # - The workers could be killed while evaluating a work item, which could # be bad if the callable being evaluated has external side-effects e.g. # writing to a file. # # To work around this problem, an exit handler is installed which tells the # workers to exit when their work queues are empty and then waits until the # threads/processes finish. _threads_wakeups = weakref.WeakKeyDictionary() _global_shutdown = False class _ThreadWakeup: def __init__(self): self._reader, self._writer = mp.Pipe(duplex=False) def close(self): self._writer.close() self._reader.close() def wakeup(self): self._writer.send_bytes(b"") def clear(self): while self._reader.poll(): self._reader.recv_bytes() def _python_exit(): global _global_shutdown _global_shutdown = True items = list(_threads_wakeups.items()) for _, thread_wakeup in items: thread_wakeup.wakeup() for t, _ in items: t.join() # Controls how many more calls than processes will be queued in the call queue. # A smaller number will mean that processes spend more time idle waiting for # work while a larger number will make Future.cancel() succeed less frequently # (Futures in the call queue cannot be cancelled). EXTRA_QUEUED_CALLS = 1 # On Windows, WaitForMultipleObjects is used to wait for processes to finish. # It can wait on, at most, 63 objects. There is an overhead of two objects: # - the result queue reader # - the thread wakeup reader _MAX_WINDOWS_WORKERS = 63 - 2 # Hack to embed stringification of remote traceback in local traceback class _RemoteTraceback(Exception): def __init__(self, tb): self.tb = tb def __str__(self): return self.tb class _ExceptionWithTraceback: def __init__(self, exc, tb): tb = traceback.format_exception(type(exc), exc, tb) tb = ''.join(tb) self.exc = exc self.tb = '\n"""\n%s"""' % tb def __reduce__(self): return _rebuild_exc, (self.exc, self.tb) def _rebuild_exc(exc, tb): exc.__cause__ = _RemoteTraceback(tb) return exc class _WorkItem(object): def __init__(self, future, fn, args, kwargs): self.future = future self.fn = fn self.args = args self.kwargs = kwargs class _ResultItem(object): def __init__(self, work_id, exception=None, result=None): self.work_id = work_id self.exception = exception self.result = result class _CallItem(object): def __init__(self, work_id, fn, args, kwargs): self.work_id = work_id self.fn = fn self.args = args self.kwargs = kwargs class _SafeQueue(Queue): """Safe Queue set exception to the future object linked to a job""" def __init__(self, max_size=0, *, ctx, pending_work_items): self.pending_work_items = pending_work_items super().__init__(max_size, ctx=ctx) def _on_queue_feeder_error(self, e, obj): if isinstance(obj, _CallItem): tb = traceback.format_exception(type(e), e, e.__traceback__) e.__cause__ = _RemoteTraceback('\n"""\n{}"""'.format(''.join(tb))) work_item = self.pending_work_items.pop(obj.work_id, None) # work_item can be None if another process terminated. In this case, # the queue_manager_thread fails all work_items with BrokenProcessPool if work_item is not None: work_item.future.set_exception(e) else: super()._on_queue_feeder_error(e, obj) def _get_chunks(*iterables, chunksize): """ Iterates over zip()ed iterables in chunks. """ it = zip(*iterables) while True: chunk = tuple(itertools.islice(it, chunksize)) if not chunk: return yield chunk def _process_chunk(fn, chunk): """ Processes a chunk of an iterable passed to map. Runs the function passed to map() on a chunk of the iterable passed to map. This function is run in a separate process. """ return [fn(*args) for args in chunk] def _sendback_result(result_queue, work_id, result=None, exception=None): """Safely send back the given result or exception""" try: result_queue.put(_ResultItem(work_id, result=result, exception=exception)) except BaseException as e: exc = _ExceptionWithTraceback(e, e.__traceback__) result_queue.put(_ResultItem(work_id, exception=exc)) def _process_worker(call_queue, result_queue, initializer, initargs): """Evaluates calls from call_queue and places the results in result_queue. This worker is run in a separate process. Args: call_queue: A ctx.Queue of _CallItems that will be read and evaluated by the worker. result_queue: A ctx.Queue of _ResultItems that will written to by the worker. initializer: A callable initializer, or None initargs: A tuple of args for the initializer """ if initializer is not None: try: initializer(*initargs) except BaseException: _base.LOGGER.critical('Exception in initializer:', exc_info=True) # The parent will notice that the process stopped and # mark the pool broken return while True: call_item = call_queue.get(block=True) if call_item is None: # Wake up queue management thread result_queue.put(os.getpid()) return try: r = call_item.fn(*call_item.args, **call_item.kwargs) except BaseException as e: exc = _ExceptionWithTraceback(e, e.__traceback__) _sendback_result(result_queue, call_item.work_id, exception=exc) else: _sendback_result(result_queue, call_item.work_id, result=r) # Liberate the resource as soon as possible, to avoid holding onto # open files or shared memory that is not needed anymore del call_item def _add_call_item_to_queue(pending_work_items, work_ids, call_queue): """Fills call_queue with _WorkItems from pending_work_items. This function never blocks. Args: pending_work_items: A dict mapping work ids to _WorkItems e.g. {5: <_WorkItem...>, 6: <_WorkItem...>, ...} work_ids: A queue.Queue of work ids e.g. Queue([5, 6, ...]). Work ids are consumed and the corresponding _WorkItems from pending_work_items are transformed into _CallItems and put in call_queue. call_queue: A multiprocessing.Queue that will be filled with _CallItems derived from _WorkItems. """ while True: if call_queue.full(): return try: work_id = work_ids.get(block=False) except queue.Empty: return else: work_item = pending_work_items[work_id] if work_item.future.set_running_or_notify_cancel(): call_queue.put(_CallItem(work_id, work_item.fn, work_item.args, work_item.kwargs), block=True) else: del pending_work_items[work_id] continue def _queue_management_worker(executor_reference, processes, pending_work_items, work_ids_queue, call_queue, result_queue, thread_wakeup): """Manages the communication between this process and the worker processes. This function is run in a local thread. Args: executor_reference: A weakref.ref to the ProcessPoolExecutor that owns this thread. Used to determine if the ProcessPoolExecutor has been garbage collected and that this function can exit. process: A list of the ctx.Process instances used as workers. pending_work_items: A dict mapping work ids to _WorkItems e.g. {5: <_WorkItem...>, 6: <_WorkItem...>, ...} work_ids_queue: A queue.Queue of work ids e.g. Queue([5, 6, ...]). call_queue: A ctx.Queue that will be filled with _CallItems derived from _WorkItems for processing by the process workers. result_queue: A ctx.SimpleQueue of _ResultItems generated by the process workers. thread_wakeup: A _ThreadWakeup to allow waking up the queue_manager_thread from the main Thread and avoid deadlocks caused by permanently locked queues. """ executor = None def shutting_down(): return (_global_shutdown or executor is None or executor._shutdown_thread) def shutdown_worker(): # This is an upper bound on the number of children alive. n_children_alive = sum(p.is_alive() for p in processes.values()) n_children_to_stop = n_children_alive n_sentinels_sent = 0 # Send the right number of sentinels, to make sure all children are # properly terminated. while n_sentinels_sent < n_children_to_stop and n_children_alive > 0: for i in range(n_children_to_stop - n_sentinels_sent): try: call_queue.put_nowait(None) n_sentinels_sent += 1 except Full: break n_children_alive = sum(p.is_alive() for p in processes.values()) # Release the queue's resources as soon as possible. call_queue.close() # If .join() is not called on the created processes then # some ctx.Queue methods may deadlock on Mac OS X. for p in processes.values(): p.join() result_reader = result_queue._reader wakeup_reader = thread_wakeup._reader readers = [result_reader, wakeup_reader] while True: _add_call_item_to_queue(pending_work_items, work_ids_queue, call_queue) # Wait for a result to be ready in the result_queue while checking # that all worker processes are still running, or for a wake up # signal send. The wake up signals come either from new tasks being # submitted, from the executor being shutdown/gc-ed, or from the # shutdown of the python interpreter. worker_sentinels = [p.sentinel for p in processes.values()] ready = wait(readers + worker_sentinels) cause = None is_broken = True if result_reader in ready: try: result_item = result_reader.recv() is_broken = False except BaseException as e: cause = traceback.format_exception(type(e), e, e.__traceback__) elif wakeup_reader in ready: is_broken = False result_item = None thread_wakeup.clear() if is_broken: # Mark the process pool broken so that submits fail right now. executor = executor_reference() if executor is not None: executor._broken = ('A child process terminated ' 'abruptly, the process pool is not ' 'usable anymore') executor._shutdown_thread = True executor = None bpe = BrokenProcessPool("A process in the process pool was " "terminated abruptly while the future was " "running or pending.") if cause is not None: bpe.__cause__ = _RemoteTraceback( f"\n'''\n{''.join(cause)}'''") # All futures in flight must be marked failed for work_id, work_item in pending_work_items.items(): work_item.future.set_exception(bpe) # Delete references to object. See issue16284 del work_item pending_work_items.clear() # Terminate remaining workers forcibly: the queues or their # locks may be in a dirty state and block forever. for p in processes.values(): p.terminate() shutdown_worker() return if isinstance(result_item, int): # Clean shutdown of a worker using its PID # (avoids marking the executor broken) assert shutting_down() p = processes.pop(result_item) p.join() if not processes: shutdown_worker() return elif result_item is not None: work_item = pending_work_items.pop(result_item.work_id, None) # work_item can be None if another process terminated (see above) if work_item is not None: if result_item.exception: work_item.future.set_exception(result_item.exception) else: work_item.future.set_result(result_item.result) # Delete references to object. See issue16284 del work_item # Delete reference to result_item del result_item # Check whether we should start shutting down. executor = executor_reference() # No more work items can be added if: # - The interpreter is shutting down OR # - The executor that owns this worker has been collected OR # - The executor that owns this worker has been shutdown. if shutting_down(): try: # Flag the executor as shutting down as early as possible if it # is not gc-ed yet. if executor is not None: executor._shutdown_thread = True # Since no new work items can be added, it is safe to shutdown # this thread if there are no pending work items. if not pending_work_items: shutdown_worker() return except Full: # This is not a problem: we will eventually be woken up (in # result_queue.get()) and be able to send a sentinel again. pass executor = None _system_limits_checked = False _system_limited = None def _check_system_limits(): global _system_limits_checked, _system_limited if _system_limits_checked: if _system_limited: raise NotImplementedError(_system_limited) _system_limits_checked = True try: nsems_max = os.sysconf("SC_SEM_NSEMS_MAX") except (AttributeError, ValueError): # sysconf not available or setting not available return if nsems_max == -1: # indetermined limit, assume that limit is determined # by available memory only return if nsems_max >= 256: # minimum number of semaphores available # according to POSIX return _system_limited = ("system provides too few semaphores (%d" " available, 256 necessary)" % nsems_max) raise NotImplementedError(_system_limited) def _chain_from_iterable_of_lists(iterable): """ Specialized implementation of itertools.chain.from_iterable. Each item in *iterable* should be a list. This function is careful not to keep references to yielded objects. """ for element in iterable: element.reverse() while element: yield element.pop() class BrokenProcessPool(_base.BrokenExecutor): """ Raised when a process in a ProcessPoolExecutor terminated abruptly while a future was in the running state. """ class ProcessPoolExecutor(_base.Executor): def __init__(self, max_workers=None, mp_context=None, initializer=None, initargs=()): """Initializes a new ProcessPoolExecutor instance. Args: max_workers: The maximum number of processes that can be used to execute the given calls. If None or not given then as many worker processes will be created as the machine has processors. mp_context: A multiprocessing context to launch the workers. This object should provide SimpleQueue, Queue and Process. initializer: A callable used to initialize worker processes. initargs: A tuple of arguments to pass to the initializer. """ _check_system_limits() if max_workers is None: self._max_workers = os.cpu_count() or 1 if sys.platform == 'win32': self._max_workers = min(_MAX_WINDOWS_WORKERS, self._max_workers) else: if max_workers <= 0: raise ValueError("max_workers must be greater than 0") elif (sys.platform == 'win32' and max_workers > _MAX_WINDOWS_WORKERS): raise ValueError( f"max_workers must be <= {_MAX_WINDOWS_WORKERS}") self._max_workers = max_workers if mp_context is None: mp_context = mp.get_context() self._mp_context = mp_context if initializer is not None and not callable(initializer): raise TypeError("initializer must be a callable") self._initializer = initializer self._initargs = initargs # Management thread self._queue_management_thread = None # Map of pids to processes self._processes = {} # Shutdown is a two-step process. self._shutdown_thread = False self._shutdown_lock = threading.Lock() self._broken = False self._queue_count = 0 self._pending_work_items = {} # Create communication channels for the executor # Make the call queue slightly larger than the number of processes to # prevent the worker processes from idling. But don't make it too big # because futures in the call queue cannot be cancelled. queue_size = self._max_workers + EXTRA_QUEUED_CALLS self._call_queue = _SafeQueue( max_size=queue_size, ctx=self._mp_context, pending_work_items=self._pending_work_items) # Killed worker processes can produce spurious "broken pipe" # tracebacks in the queue's own worker thread. But we detect killed # processes anyway, so silence the tracebacks. self._call_queue._ignore_epipe = True self._result_queue = mp_context.SimpleQueue() self._work_ids = queue.Queue() # _ThreadWakeup is a communication channel used to interrupt the wait # of the main loop of queue_manager_thread from another thread (e.g. # when calling executor.submit or executor.shutdown). We do not use the # _result_queue to send the wakeup signal to the queue_manager_thread # as it could result in a deadlock if a worker process dies with the # _result_queue write lock still acquired. self._queue_management_thread_wakeup = _ThreadWakeup() def _start_queue_management_thread(self): if self._queue_management_thread is None: # When the executor gets garbarge collected, the weakref callback # will wake up the queue management thread so that it can terminate # if there is no pending work item. def weakref_cb(_, thread_wakeup=self._queue_management_thread_wakeup): mp.util.debug('Executor collected: triggering callback for' ' QueueManager wakeup') thread_wakeup.wakeup() # Start the processes so that their sentinels are known. self._adjust_process_count() self._queue_management_thread = threading.Thread( target=_queue_management_worker, args=(weakref.ref(self, weakref_cb), self._processes, self._pending_work_items, self._work_ids, self._call_queue, self._result_queue, self._queue_management_thread_wakeup), name="QueueManagerThread") self._queue_management_thread.daemon = True self._queue_management_thread.start() _threads_wakeups[self._queue_management_thread] = \ self._queue_management_thread_wakeup def _adjust_process_count(self): for _ in range(len(self._processes), self._max_workers): p = self._mp_context.Process( target=_process_worker, args=(self._call_queue, self._result_queue, self._initializer, self._initargs)) p.start() self._processes[p.pid] = p def submit(*args, **kwargs): if len(args) >= 2: self, fn, *args = args elif not args: raise TypeError("descriptor 'submit' of 'ProcessPoolExecutor' object " "needs an argument") elif 'fn' in kwargs: fn = kwargs.pop('fn') self, *args = args else: raise TypeError('submit expected at least 1 positional argument, ' 'got %d' % (len(args)-1)) with self._shutdown_lock: if self._broken: raise BrokenProcessPool(self._broken) if self._shutdown_thread: raise RuntimeError('cannot schedule new futures after shutdown') if _global_shutdown: raise RuntimeError('cannot schedule new futures after ' 'interpreter shutdown') f = _base.Future() w = _WorkItem(f, fn, args, kwargs) self._pending_work_items[self._queue_count] = w self._work_ids.put(self._queue_count) self._queue_count += 1 # Wake up queue management thread self._queue_management_thread_wakeup.wakeup() self._start_queue_management_thread() return f submit.__doc__ = _base.Executor.submit.__doc__ def map(self, fn, *iterables, timeout=None, chunksize=1): """Returns an iterator equivalent to map(fn, iter). Args: fn: A callable that will take as many arguments as there are passed iterables. timeout: The maximum number of seconds to wait. If None, then there is no limit on the wait time. chunksize: If greater than one, the iterables will be chopped into chunks of size chunksize and submitted to the process pool. If set to one, the items in the list will be sent one at a time. Returns: An iterator equivalent to: map(func, *iterables) but the calls may be evaluated out-of-order. Raises: TimeoutError: If the entire result iterator could not be generated before the given timeout. Exception: If fn(*args) raises for any values. """ if chunksize < 1: raise ValueError("chunksize must be >= 1.") results = super().map(partial(_process_chunk, fn), _get_chunks(*iterables, chunksize=chunksize), timeout=timeout) return _chain_from_iterable_of_lists(results) def shutdown(self, wait=True): with self._shutdown_lock: self._shutdown_thread = True if self._queue_management_thread: # Wake up queue management thread self._queue_management_thread_wakeup.wakeup() if wait: self._queue_management_thread.join() # To reduce the risk of opening too many files, remove references to # objects that use file descriptors. self._queue_management_thread = None if self._call_queue is not None: self._call_queue.close() if wait: self._call_queue.join_thread() self._call_queue = None self._result_queue = None self._processes = None if self._queue_management_thread_wakeup: self._queue_management_thread_wakeup.close() self._queue_management_thread_wakeup = None shutdown.__doc__ = _base.Executor.shutdown.__doc__ atexit.register(_python_exit)