error： _multiprocessing.SemLock( FileNotFoundError: [Errno 2] No such file or directory

[TYTaO]

I successfully ran the pytorch example. But when i use datalaoder prefetch_factor in pytorch example.
like this

train_loader = torch.utils.data.DataLoader(
        datasets.MNIST('data', train=True, download=True,
                       transform=transforms.Compose([
                           transforms.ToTensor(),
                           transforms.Normalize((0.1307,), (0.3081,))
                       ])),
        batch_size=BATCH_SIZE, shuffle=True, num_workers=2, prefetch_factor=2)

i got this error:

Traceback (most recent call last):
  File "//./pytorchexample.py", line 112, in <module>
    train(model, DEVICE, train_loader, optimizer, epoch)
  File "//./pytorchexample.py", line 49, in train
    for batch_idx, (data, target) in enumerate(train_loader):
  File "/usr/local/lib/python3.9/dist-packages/torch/utils/data/dataloader.py", line 368, in __iter__
    return self._get_iterator()
  File "/usr/local/lib/python3.9/dist-packages/torch/utils/data/dataloader.py", line 314, in _get_iterator
    return _MultiProcessingDataLoaderIter(self)
  File "/usr/local/lib/python3.9/dist-packages/torch/utils/data/dataloader.py", line 900, in __init__
    self._worker_result_queue = multiprocessing_context.Queue()  # type: ignore[var-annotated]
  File "/usr/lib/python3.9/multiprocessing/context.py", line 103, in Queue
    return Queue(maxsize, ctx=self.get_context())
  File "/usr/lib/python3.9/multiprocessing/queues.py", line 43, in __init__
    self._rlock = ctx.Lock()
  File "/usr/lib/python3.9/multiprocessing/context.py", line 68, in Lock
    return Lock(ctx=self.get_context())
  File "/usr/lib/python3.9/multiprocessing/synchronize.py", line 162, in __init__
    SemLock.__init__(self, SEMAPHORE, 1, 1, ctx=ctx)
  File "/usr/lib/python3.9/multiprocessing/synchronize.py", line 57, in __init__
    sl = self._semlock = _multiprocessing.SemLock(
FileNotFoundError: [Errno 2] No such file or directory

[dimakuv]

Duplicate of gramineproject/graphene#2689

Unfortunately, it is a known issue -- Gramine doesn't support the multiprocessing package of Python. This is because Gramine currently doesn't support Sys-V semaphores, which the multiprocessing package requires.

UPDATE 24. March 2023. Python's multiprocessing package uses POSIX semaphores (and thus shared memory) and not Sys-V semaphores. See:

• multiprocessing: use SysV semaphores on FreeBSD python/cpython#54557
• https://github.com/python/cpython/blob/a87c46eab3c306b1c5b8a072b7b30ac2c50651c0/Modules/_multiprocessing/semaphore.c#L210-L219

That's unfortunate, because implementing POSIX semaphores in Gramine/SGX would require allowing untrusted shared memory (/dev/shm), which will probably never happen...