3 from __future__ import annotations
5 from abc import ABC, abstractmethod
6 import concurrent.futures as fut
7 from collections import defaultdict
8 from dataclasses import dataclass
17 from typing import Any, Callable, Dict, List, Optional, Set
19 import cloudpickle # type: ignore
20 from overrides import overrides
22 from ansi import bg, fg, underline, reset
25 from exec_utils import run_silently, cmd_in_background, cmd_with_timeout
26 from decorator_utils import singleton
27 import histogram as hist
29 logger = logging.getLogger(__name__)
31 parser = config.add_commandline_args(
32 f"Executors ({__file__})",
33 "Args related to processing executors."
36 '--executors_threadpool_size',
39 help='Number of threads in the default threadpool, leave unset for default',
43 '--executors_processpool_size',
46 help='Number of processes in the default processpool, leave unset for default',
50 '--executors_schedule_remote_backups',
52 action=argparse_utils.ActionNoYes,
53 help='Should we schedule duplicative backup work if a remote bundle is slow',
56 '--executors_max_bundle_failures',
60 help='Maximum number of failures before giving up on a bundle',
63 RSYNC = 'rsync -q --no-motd -W --ignore-existing --timeout=60 --size-only -z'
64 SSH = 'ssh -oForwardX11=no'
67 def make_cloud_pickle(fun, *args, **kwargs):
68 logger.debug(f"Making cloudpickled bundle at {fun.__name__}")
69 return cloudpickle.dumps((fun, args, kwargs))
72 class BaseExecutor(ABC):
73 def __init__(self, *, title=''):
76 self.histogram = hist.SimpleHistogram(
77 hist.SimpleHistogram.n_evenly_spaced_buckets(
86 **kwargs) -> fut.Future:
91 wait: bool = True) -> None:
94 def adjust_task_count(self, delta: int) -> None:
95 self.task_count += delta
96 logger.debug(f'Executor current task count is {self.task_count}')
99 class ThreadExecutor(BaseExecutor):
101 max_workers: Optional[int] = None):
104 if max_workers is not None:
105 workers = max_workers
106 elif 'executors_threadpool_size' in config.config:
107 workers = config.config['executors_threadpool_size']
108 logger.debug(f'Creating threadpool executor with {workers} workers')
109 self._thread_pool_executor = fut.ThreadPoolExecutor(
111 thread_name_prefix="thread_executor_helper"
114 def run_local_bundle(self, fun, *args, **kwargs):
115 logger.debug(f"Running local bundle at {fun.__name__}")
117 result = fun(*args, **kwargs)
119 self.adjust_task_count(-1)
120 duration = end - start
121 logger.debug(f"{fun.__name__} finished; used {duration:.1f}s")
122 self.histogram.add_item(duration)
129 **kwargs) -> fut.Future:
130 self.adjust_task_count(+1)
132 newargs.append(function)
135 return self._thread_pool_executor.submit(
136 self.run_local_bundle,
142 wait = True) -> None:
143 logger.debug(f'Shutting down threadpool executor {self.title}')
144 print(self.histogram)
145 self._thread_pool_executor.shutdown(wait)
148 class ProcessExecutor(BaseExecutor):
153 if max_workers is not None:
154 workers = max_workers
155 elif 'executors_processpool_size' in config.config:
156 workers = config.config['executors_processpool_size']
157 logger.debug(f'Creating processpool executor with {workers} workers.')
158 self._process_executor = fut.ProcessPoolExecutor(
162 def run_cloud_pickle(self, pickle):
163 fun, args, kwargs = cloudpickle.loads(pickle)
164 logger.debug(f"Running pickled bundle at {fun.__name__}")
165 result = fun(*args, **kwargs)
166 self.adjust_task_count(-1)
173 **kwargs) -> fut.Future:
175 self.adjust_task_count(+1)
176 pickle = make_cloud_pickle(function, *args, **kwargs)
177 result = self._process_executor.submit(
178 self.run_cloud_pickle,
181 result.add_done_callback(
182 lambda _: self.histogram.add_item(
189 def shutdown(self, wait=True) -> None:
190 logger.debug(f'Shutting down processpool executor {self.title}')
191 self._process_executor.shutdown(wait)
192 print(self.histogram)
194 def __getstate__(self):
195 state = self.__dict__.copy()
196 state['_process_executor'] = None
200 class RemoteExecutorException(Exception):
201 """Thrown when a bundle cannot be executed despite several retries."""
206 class RemoteWorkerRecord:
213 return hash((self.username, self.machine))
216 return f'{self.username}@{self.machine}'
224 worker: Optional[RemoteWorkerRecord]
225 username: Optional[str]
226 machine: Optional[str]
235 src_bundle: BundleDetails
236 is_cancelled: threading.Event
238 backup_bundles: Optional[List[BundleDetails]]
243 if uuid[-9:-2] == '_backup':
245 suffix = f'{uuid[-6:]}_b{self.uuid[-1:]}'
255 fg('marigold yellow'),
258 fg('cornflower blue'),
259 fg('turquoise blue'),
261 fg('lavender purple'),
264 c = colorz[int(uuid[-2:], 16) % len(colorz)]
265 fname = self.fname if self.fname is not None else 'nofname'
266 machine = self.machine if self.machine is not None else 'nomachine'
267 return f'{c}{suffix}/{fname}/{machine}{reset()}'
270 class RemoteExecutorStatus:
271 def __init__(self, total_worker_count: int) -> None:
272 self.worker_count = total_worker_count
273 self.known_workers: Set[RemoteWorkerRecord] = set()
274 self.start_per_bundle: Dict[str, float] = defaultdict(float)
275 self.end_per_bundle: Dict[str, float] = defaultdict(float)
276 self.finished_bundle_timings_per_worker: Dict[
280 self.in_flight_bundles_by_worker: Dict[
284 self.bundle_details_by_uuid: Dict[str, BundleDetails] = {}
285 self.finished_bundle_timings: List[float] = []
286 self.last_periodic_dump: Optional[float] = None
287 self.total_bundles_submitted = 0
289 # Protects reads and modification using self. Also used
290 # as a memory fence for modifications to bundle.
291 self.lock = threading.Lock()
293 def record_acquire_worker(
295 worker: RemoteWorkerRecord,
299 self.record_acquire_worker_already_locked(
304 def record_acquire_worker_already_locked(
306 worker: RemoteWorkerRecord,
309 assert self.lock.locked()
310 self.known_workers.add(worker)
311 self.start_per_bundle[uuid] = None
312 x = self.in_flight_bundles_by_worker.get(worker, set())
314 self.in_flight_bundles_by_worker[worker] = x
316 def record_bundle_details(
318 details: BundleDetails) -> None:
320 self.record_bundle_details_already_locked(details)
322 def record_bundle_details_already_locked(
324 details: BundleDetails) -> None:
325 assert self.lock.locked()
326 self.bundle_details_by_uuid[details.uuid] = details
328 def record_release_worker(
330 worker: RemoteWorkerRecord,
335 self.record_release_worker_already_locked(
336 worker, uuid, was_cancelled
339 def record_release_worker_already_locked(
341 worker: RemoteWorkerRecord,
345 assert self.lock.locked()
347 self.end_per_bundle[uuid] = ts
348 self.in_flight_bundles_by_worker[worker].remove(uuid)
349 if not was_cancelled:
350 bundle_latency = ts - self.start_per_bundle[uuid]
351 x = self.finished_bundle_timings_per_worker.get(worker, list())
352 x.append(bundle_latency)
353 self.finished_bundle_timings_per_worker[worker] = x
354 self.finished_bundle_timings.append(bundle_latency)
356 def record_processing_began(self, uuid: str):
358 self.start_per_bundle[uuid] = time.time()
360 def total_in_flight(self) -> int:
361 assert self.lock.locked()
363 for worker in self.known_workers:
364 total_in_flight += len(self.in_flight_bundles_by_worker[worker])
365 return total_in_flight
367 def total_idle(self) -> int:
368 assert self.lock.locked()
369 return self.worker_count - self.total_in_flight()
372 assert self.lock.locked()
374 total_finished = len(self.finished_bundle_timings)
375 total_in_flight = self.total_in_flight()
376 ret = f'\n\n{underline()}Remote Executor Pool Status{reset()}: '
378 if len(self.finished_bundle_timings) > 1:
379 qall = numpy.quantile(self.finished_bundle_timings, [0.5, 0.95])
381 f'⏱=∀p50:{qall[0]:.1f}s, ∀p95:{qall[1]:.1f}s, '
382 f'✅={total_finished}/{self.total_bundles_submitted}, '
383 f'💻n={total_in_flight}/{self.worker_count}\n'
387 f' ✅={total_finished}/{self.total_bundles_submitted}, '
388 f'💻n={total_in_flight}/{self.worker_count}\n'
391 for worker in self.known_workers:
392 ret += f' {fg("lightning yellow")}{worker.machine}{reset()}: '
393 timings = self.finished_bundle_timings_per_worker.get(worker, [])
397 qworker = numpy.quantile(timings, [0.5, 0.95])
398 ret += f' 💻p50: {qworker[0]:.1f}s, 💻p95: {qworker[1]:.1f}s\n'
402 ret += f' ...finished {count} total bundle(s) so far\n'
403 in_flight = len(self.in_flight_bundles_by_worker[worker])
405 ret += f' ...{in_flight} bundles currently in flight:\n'
406 for bundle_uuid in self.in_flight_bundles_by_worker[worker]:
407 details = self.bundle_details_by_uuid.get(
411 pid = str(details.pid) if (details and details.pid != 0) else "TBD"
412 if self.start_per_bundle[bundle_uuid] is not None:
413 sec = ts - self.start_per_bundle[bundle_uuid]
414 ret += f' (pid={pid}): {details} for {sec:.1f}s so far '
416 ret += f' {details} setting up / copying data...'
419 if qworker is not None:
421 ret += f'{bg("red")}>💻p95{reset()} '
422 elif sec > qworker[0]:
423 ret += f'{fg("red")}>💻p50{reset()} '
425 if sec > qall[1] * 1.5:
426 ret += f'{bg("red")}!!!{reset()}'
427 if details is not None:
428 logger.debug(f'Flagging {details} for another backup')
429 details.super_slow = True
431 ret += f'{bg("red")}>∀p95{reset()} '
432 if details is not None:
433 logger.debug(f'Flagging {details} for a backup')
434 details.too_slow = True
436 ret += f'{fg("red")}>∀p50{reset()}'
440 def periodic_dump(self, total_bundles_submitted: int) -> None:
441 assert self.lock.locked()
442 self.total_bundles_submitted = total_bundles_submitted
445 self.last_periodic_dump is None
446 or ts - self.last_periodic_dump > 5.0
449 self.last_periodic_dump = ts
452 class RemoteWorkerSelectionPolicy(ABC):
453 def register_worker_pool(self, workers):
455 self.workers = workers
458 def is_worker_available(self) -> bool:
464 machine_to_avoid = None
465 ) -> Optional[RemoteWorkerRecord]:
469 class WeightedRandomRemoteWorkerSelectionPolicy(RemoteWorkerSelectionPolicy):
470 def is_worker_available(self) -> bool:
471 for worker in self.workers:
478 machine_to_avoid = None
479 ) -> Optional[RemoteWorkerRecord]:
481 for worker in self.workers:
482 for x in range(0, worker.count):
483 for y in range(0, worker.weight):
484 grabbag.append(worker)
486 for _ in range(0, 5):
487 random.shuffle(grabbag)
489 if worker.machine != machine_to_avoid or _ > 2:
492 logger.debug(f'Selected worker {worker}')
494 logger.warning("Couldn't find a worker; go fish.")
498 class RoundRobinRemoteWorkerSelectionPolicy(RemoteWorkerSelectionPolicy):
499 def __init__(self) -> None:
502 def is_worker_available(self) -> bool:
503 for worker in self.workers:
510 machine_to_avoid: str = None
511 ) -> Optional[RemoteWorkerRecord]:
514 worker = self.workers[x]
518 if x >= len(self.workers):
521 logger.debug(f'Selected worker {worker}')
524 if x >= len(self.workers):
527 logger.warning("Couldn't find a worker; go fish.")
531 class RemoteExecutor(BaseExecutor):
533 workers: List[RemoteWorkerRecord],
534 policy: RemoteWorkerSelectionPolicy) -> None:
536 self.workers = workers
538 self.worker_count = 0
539 for worker in self.workers:
540 self.worker_count += worker.count
541 if self.worker_count <= 0:
542 msg = f"We need somewhere to schedule work; count was {self.worker_count}"
544 raise RemoteExecutorException(msg)
545 self.policy.register_worker_pool(self.workers)
546 self.cv = threading.Condition()
547 logger.debug(f'Creating {self.worker_count} local threads, one per remote worker.')
548 self._helper_executor = fut.ThreadPoolExecutor(
549 thread_name_prefix="remote_executor_helper",
550 max_workers=self.worker_count,
552 self.status = RemoteExecutorStatus(self.worker_count)
553 self.total_bundles_submitted = 0
555 def is_worker_available(self) -> bool:
556 return self.policy.is_worker_available()
560 machine_to_avoid: str = None
561 ) -> Optional[RemoteWorkerRecord]:
562 return self.policy.acquire_worker(machine_to_avoid)
564 def find_available_worker_or_block(
566 machine_to_avoid: str = None
567 ) -> RemoteWorkerRecord:
569 while not self.is_worker_available():
571 worker = self.acquire_worker(machine_to_avoid)
572 if worker is not None:
574 msg = "We should never reach this point in the code"
578 def release_worker(self, worker: RemoteWorkerRecord) -> None:
579 logger.debug(f'Released worker {worker}')
584 def heartbeat(self) -> None:
585 with self.status.lock:
586 # Regular progress report
587 self.status.periodic_dump(self.total_bundles_submitted)
589 # Look for bundles to reschedule.
590 num_done = len(self.status.finished_bundle_timings)
591 if num_done > 7 or (num_done > 5 and self.is_worker_available()):
592 for worker, bundle_uuids in self.status.in_flight_bundles_by_worker.items():
593 for uuid in bundle_uuids:
594 bundle = self.status.bundle_details_by_uuid.get(uuid, None)
596 bundle is not None and
598 bundle.src_bundle is None and
599 config.config['executors_schedule_remote_backups']
601 self.consider_backup_for_bundle(bundle)
603 def consider_backup_for_bundle(self, bundle: BundleDetails) -> None:
604 assert self.status.lock.locked()
607 and len(bundle.backup_bundles) == 0 # one backup per
609 msg = f"*** Rescheduling {bundle} (first backup) ***"
611 self.schedule_backup_for_bundle(bundle)
615 and len(bundle.backup_bundles) < 2 # two backups in dire situations
616 and self.status.total_idle() > 4
618 msg = f"*** Rescheduling {bundle} (second backup) ***"
620 self.schedule_backup_for_bundle(bundle)
623 def check_if_cancelled(self, bundle: BundleDetails) -> bool:
624 with self.status.lock:
625 if bundle.is_cancelled.wait(timeout=0.0):
626 logger.debug(f'Bundle {bundle.uuid} is cancelled, bail out.')
627 bundle.was_cancelled = True
631 def launch(self, bundle: BundleDetails, override_avoid_machine=None) -> Any:
632 """Find a worker for bundle or block until one is available."""
633 self.adjust_task_count(+1)
635 hostname = bundle.hostname
636 avoid_machine = override_avoid_machine
637 is_original = bundle.src_bundle is None
639 # Try not to schedule a backup on the same host as the original.
640 if avoid_machine is None and bundle.src_bundle is not None:
641 avoid_machine = bundle.src_bundle.machine
643 while worker is None:
644 worker = self.find_available_worker_or_block(avoid_machine)
646 # Ok, found a worker.
647 bundle.worker = worker
648 machine = bundle.machine = worker.machine
649 username = bundle.username = worker.username
651 self.status.record_acquire_worker(worker, uuid)
652 logger.debug(f'{bundle}: Running bundle on {worker}...')
654 # Before we do any work, make sure the bundle is still viable.
655 if self.check_if_cancelled(bundle):
657 return self.post_launch_work(bundle)
658 except Exception as e:
661 f'{bundle}: bundle says it\'s cancelled upfront but no results?!'
663 assert bundle.worker is not None
664 self.status.record_release_worker(
669 self.release_worker(bundle.worker)
670 self.adjust_task_count(-1)
672 # Weird. We are the original owner of this
673 # bundle. For it to have been cancelled, a backup
674 # must have already started and completed before
675 # we even for started. Moreover, the backup says
676 # it is done but we can't find the results it
677 # should have copied over. Reschedule the whole
679 return self.emergency_retry_nasty_bundle(bundle)
681 # Expected(?). We're a backup and our bundle is
682 # cancelled before we even got started. Something
683 # went bad in post_launch_work (I acutually don't
684 # see what?) but probably not worth worrying
688 # Send input code / data to worker machine if it's not local.
689 if hostname not in machine:
691 cmd = f'{RSYNC} {bundle.code_file} {username}@{machine}:{bundle.code_file}'
692 start_ts = time.time()
693 logger.info(f"{bundle}: Copying work to {worker} via {cmd}.")
695 xfer_latency = time.time() - start_ts
696 logger.info(f"{bundle}: Copying done to {worker} in {xfer_latency:.1f}s.")
697 except Exception as e:
700 f'{bundle}: failed to send instructions to worker machine?!?'
702 assert bundle.worker is not None
703 self.status.record_release_worker(
708 self.release_worker(bundle.worker)
709 self.adjust_task_count(-1)
711 # Weird. We tried to copy the code to the worker and it failed...
712 # And we're the original bundle. We have to retry.
713 return self.emergency_retry_nasty_bundle(bundle)
715 # This is actually expected; we're a backup.
716 # There's a race condition where someone else
717 # already finished the work and removed the source
718 # code file before we could copy it. No biggie.
721 # Kick off the work. Note that if this fails we let
722 # wait_for_process deal with it.
723 self.status.record_processing_began(uuid)
724 cmd = (f'{SSH} {bundle.username}@{bundle.machine} '
725 f'"source py39-venv/bin/activate &&'
726 f' /home/scott/lib/python_modules/remote_worker.py'
727 f' --code_file {bundle.code_file} --result_file {bundle.result_file}"')
728 logger.debug(f'{bundle}: Executing {cmd} in the background to kick off work...')
729 p = cmd_in_background(cmd, silent=True)
730 bundle.pid = pid = p.pid
731 logger.debug(f'{bundle}: Local ssh process pid={pid}; remote worker is {machine}.')
732 return self.wait_for_process(p, bundle, 0)
734 def wait_for_process(self, p: subprocess.Popen, bundle: BundleDetails, depth: int) -> Any:
735 machine = bundle.machine
739 f"I've gotten repeated errors waiting on this bundle; giving up on pid={pid}."
742 self.status.record_release_worker(
747 self.release_worker(bundle.worker)
748 self.adjust_task_count(-1)
749 return self.emergency_retry_nasty_bundle(bundle)
751 # Spin until either the ssh job we scheduled finishes the
752 # bundle or some backup worker signals that they finished it
757 except subprocess.TimeoutExpired:
759 if self.check_if_cancelled(bundle):
761 f'{bundle}: another worker finished bundle, checking it out...'
766 f"{bundle}: pid {pid} ({machine}) our ssh finished, checking it out..."
771 # If we get here we believe the bundle is done; either the ssh
772 # subprocess finished (hopefully successfully) or we noticed
773 # that some other worker seems to have completed the bundle
774 # and we're bailing out.
776 ret = self.post_launch_work(bundle)
777 if ret is not None and p is not None:
781 # Something went wrong; e.g. we could not copy the results
782 # back, cleanup after ourselves on the remote machine, or
783 # unpickle the results we got from the remove machine. If we
784 # still have an active ssh subprocess, keep waiting on it.
785 # Otherwise, time for an emergency reschedule.
786 except Exception as e:
788 logger.error(f'{bundle}: Something unexpected just happened...')
791 f"{bundle}: Failed to wrap up \"done\" bundle, re-waiting on active ssh."
793 return self.wait_for_process(p, bundle, depth + 1)
795 self.status.record_release_worker(
800 self.release_worker(bundle.worker)
801 self.adjust_task_count(-1)
802 return self.emergency_retry_nasty_bundle(bundle)
804 def post_launch_work(self, bundle: BundleDetails) -> Any:
805 with self.status.lock:
806 is_original = bundle.src_bundle is None
807 was_cancelled = bundle.was_cancelled
808 username = bundle.username
809 machine = bundle.machine
810 result_file = bundle.result_file
811 code_file = bundle.code_file
813 # Whether original or backup, if we finished first we must
814 # fetch the results if the computation happened on a
816 bundle.end_ts = time.time()
817 if not was_cancelled:
818 assert bundle.machine is not None
819 if bundle.hostname not in bundle.machine:
820 cmd = f'{RSYNC} {username}@{machine}:{result_file} {result_file} 2>/dev/null'
822 f"{bundle}: Fetching results from {username}@{machine} via {cmd}"
825 # If either of these throw they are handled in
828 run_silently(f'{SSH} {username}@{machine}'
829 f' "/bin/rm -f {code_file} {result_file}"')
830 dur = bundle.end_ts - bundle.start_ts
831 self.histogram.add_item(dur)
833 # Only the original worker should unpickle the file contents
834 # though since it's the only one whose result matters. The
835 # original is also the only job that may delete result_file
836 # from disk. Note that the original may have been cancelled
837 # if one of the backups finished first; it still must read the
840 logger.debug(f"{bundle}: Unpickling {result_file}.")
842 with open(f'{result_file}', 'rb') as rb:
843 serialized = rb.read()
844 result = cloudpickle.loads(serialized)
845 except Exception as e:
846 msg = f'Failed to load {result_file}, this is bad news.'
848 self.status.record_release_worker(
853 self.release_worker(bundle.worker)
855 # Re-raise the exception; the code in wait_for_process may
856 # decide to emergency_retry_nasty_bundle here.
860 f'Removing local (master) {code_file} and {result_file}.'
862 os.remove(f'{result_file}')
863 os.remove(f'{code_file}')
865 # Notify any backups that the original is done so they
866 # should stop ASAP. Do this whether or not we
867 # finished first since there could be more than one
869 if bundle.backup_bundles is not None:
870 for backup in bundle.backup_bundles:
872 f'{bundle}: Notifying backup {backup.uuid} that it\'s cancelled'
874 backup.is_cancelled.set()
876 # This is a backup job and, by now, we have already fetched
877 # the bundle results.
879 # Backup results don't matter, they just need to leave the
880 # result file in the right place for their originals to
881 # read/unpickle later.
884 # Tell the original to stop if we finished first.
885 if not was_cancelled:
887 f'{bundle}: Notifying original {bundle.src_bundle.uuid} we beat them to it.'
889 bundle.src_bundle.is_cancelled.set()
891 assert bundle.worker is not None
892 self.status.record_release_worker(
897 self.release_worker(bundle.worker)
898 self.adjust_task_count(-1)
901 def create_original_bundle(self, pickle, fname: str):
902 from string_utils import generate_uuid
903 uuid = generate_uuid(as_hex=True)
904 code_file = f'/tmp/{uuid}.code.bin'
905 result_file = f'/tmp/{uuid}.result.bin'
907 logger.debug(f'Writing pickled code to {code_file}')
908 with open(f'{code_file}', 'wb') as wb:
911 bundle = BundleDetails(
912 pickled_code = pickle,
918 hostname = platform.node(),
919 code_file = code_file,
920 result_file = result_file,
922 start_ts = time.time(),
927 is_cancelled = threading.Event(),
928 was_cancelled = False,
932 self.status.record_bundle_details(bundle)
933 logger.debug(f'{bundle}: Created an original bundle')
936 def create_backup_bundle(self, src_bundle: BundleDetails):
937 assert src_bundle.backup_bundles is not None
938 n = len(src_bundle.backup_bundles)
939 uuid = src_bundle.uuid + f'_backup#{n}'
941 backup_bundle = BundleDetails(
942 pickled_code = src_bundle.pickled_code,
944 fname = src_bundle.fname,
948 hostname = src_bundle.hostname,
949 code_file = src_bundle.code_file,
950 result_file = src_bundle.result_file,
952 start_ts = time.time(),
956 src_bundle = src_bundle,
957 is_cancelled = threading.Event(),
958 was_cancelled = False,
959 backup_bundles = None, # backup backups not allowed
962 src_bundle.backup_bundles.append(backup_bundle)
963 self.status.record_bundle_details_already_locked(backup_bundle)
964 logger.debug(f'{backup_bundle}: Created a backup bundle')
967 def schedule_backup_for_bundle(self,
968 src_bundle: BundleDetails):
969 assert self.status.lock.locked()
970 backup_bundle = self.create_backup_bundle(src_bundle)
972 f'{backup_bundle.uuid}/{backup_bundle.fname}: Scheduling backup for execution...'
974 self._helper_executor.submit(self.launch, backup_bundle)
976 # Results from backups don't matter; if they finish first
977 # they will move the result_file to this machine and let
978 # the original pick them up and unpickle them.
980 def emergency_retry_nasty_bundle(self, bundle: BundleDetails) -> fut.Future:
981 is_original = bundle.src_bundle is None
983 avoid_last_machine = bundle.machine
984 bundle.machine = None
985 bundle.username = None
986 bundle.failure_count += 1
992 if bundle.failure_count > retry_limit:
994 f'{bundle}: Tried this bundle too many times already ({retry_limit}x); giving up.'
997 raise RemoteExecutorException(
998 f'{bundle}: This bundle can\'t be completed despite several backups and retries'
1001 logger.error(f'{bundle}: At least it\'s only a backup; better luck with the others.')
1005 f'>>> Emergency rescheduling {bundle} because of unexected errors (wtf?!) <<<'
1007 return self.launch(bundle, avoid_last_machine)
1013 **kwargs) -> fut.Future:
1014 pickle = make_cloud_pickle(function, *args, **kwargs)
1015 bundle = self.create_original_bundle(pickle, function.__name__)
1016 self.total_bundles_submitted += 1
1017 return self._helper_executor.submit(self.launch, bundle)
1020 def shutdown(self, wait=True) -> None:
1021 self._helper_executor.shutdown(wait)
1022 logging.debug(f'Shutting down RemoteExecutor {self.title}')
1023 print(self.histogram)
1027 class DefaultExecutors(object):
1029 self.thread_executor: Optional[ThreadExecutor] = None
1030 self.process_executor: Optional[ProcessExecutor] = None
1031 self.remote_executor: Optional[RemoteExecutor] = None
1033 def ping(self, host) -> bool:
1034 logger.debug(f'RUN> ping -c 1 {host}')
1036 x = cmd_with_timeout(
1037 f'ping -c 1 {host} >/dev/null 2>/dev/null',
1044 def thread_pool(self) -> ThreadExecutor:
1045 if self.thread_executor is None:
1046 self.thread_executor = ThreadExecutor()
1047 return self.thread_executor
1049 def process_pool(self) -> ProcessExecutor:
1050 if self.process_executor is None:
1051 self.process_executor = ProcessExecutor()
1052 return self.process_executor
1054 def remote_pool(self) -> RemoteExecutor:
1055 if self.remote_executor is None:
1056 logger.info('Looking for some helper machines...')
1057 pool: List[RemoteWorkerRecord] = []
1058 if self.ping('cheetah.house'):
1059 logger.info('Found cheetah.house')
1063 machine = 'cheetah.house',
1068 if self.ping('video.house'):
1069 logger.info('Found video.house')
1073 machine = 'video.house',
1078 if self.ping('wannabe.house'):
1079 logger.info('Found wannabe.house')
1083 machine = 'wannabe.house',
1088 if self.ping('meerkat.cabin'):
1089 logger.info('Found meerkat.cabin')
1093 machine = 'meerkat.cabin',
1098 if self.ping('backup.house'):
1099 logger.info('Found backup.house')
1103 machine = 'backup.house',
1108 if self.ping('kiosk.house'):
1109 logger.info('Found kiosk.house')
1113 machine = 'kiosk.house',
1118 if self.ping('puma.cabin'):
1119 logger.info('Found puma.cabin')
1123 machine = 'puma.cabin',
1129 # The controller machine has a lot to do; go easy on it.
1131 if record.machine == platform.node() and record.count > 1:
1132 logger.info(f'Reducing workload for {record.machine}.')
1135 policy = WeightedRandomRemoteWorkerSelectionPolicy()
1136 policy.register_worker_pool(pool)
1137 self.remote_executor = RemoteExecutor(pool, policy)
1138 return self.remote_executor