search

hail-is / hail

Cloud-native genomic dataframes and batch computing
https://hail.is
MIT License
984 stars 247 forks source link

[batch] Too many open files when submitting a batch #13940

Open danking opened 1 year ago

danking commented 1 year ago

What happened?

https://hail.zulipchat.com/#narrow/stream/223457-Hail-Batch-support/topic/Too.20many.20open.20files.20exception

Version

0.2.124

Relevant log output

Task exception was never retrieved
future: <Task finished name='Task-29959' coro=<BaseSession.post() done, defined at /usr/local/lib/python3.9/site-packages/hailtop/aiocloud/common/session.py:20> exception=ServerDisconnectedError('Server disconnected')>
Traceback (most recent call last):
  File "/usr/local/lib/python3.9/site-packages/aiohttp/connector.py", line 986, in _wrap_create_connection
    return await self._loop.create_connection(*args, **kwargs)  # type: ignore[return-value]  # noqa
  File "/usr/local/Cellar/python@3.9/3.9.17_1/Frameworks/Python.framework/Versions/3.9/lib/python3.9/asyncio/base_events.py", line 1065, in create_connection
    raise exceptions[0]
  File "/usr/local/Cellar/python@3.9/3.9.17_1/Frameworks/Python.framework/Versions/3.9/lib/python3.9/asyncio/base_events.py", line 1050, in create_connection
    sock = await self._connect_sock(
  File "/usr/local/Cellar/python@3.9/3.9.17_1/Frameworks/Python.framework/Versions/3.9/lib/python3.9/asyncio/base_events.py", line 944, in _connect_sock
    sock = socket.socket(family=family, type=type_, proto=proto)
  File "/usr/local/Cellar/python@3.9/3.9.17_1/Frameworks/Python.framework/Versions/3.9/lib/python3.9/socket.py", line 232, in __init__
    _socket.socket.__init__(self, family, type, proto, fileno)
OSError: [Errno 24] Too many open files

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/Users/weisburd/code/sma_finder/sma_finder_pipeline.py", line 473, in <module>
    main()
  File "/Users/weisburd/code/sma_finder/sma_finder_pipeline.py", line 393, in main
    bp.run()
  File "/Users/weisburd/code/step-pipeline/step_pipeline/batch.py", line 300, in run
    result = self._run_batch_obj()
  File "/Users/weisburd/code/step-pipeline/step_pipeline/batch.py", line 368, in _run_batch_obj
    result = self._batch.run(
  File "/usr/local/lib/python3.9/site-packages/hailtop/batch/batch.py", line 712, in run
    run_result = self._backend._run(self, dry_run, verbose, delete_scratch_on_exit, **backend_kwargs)  # pylint: disable=assignment-from-no-return
  File "/usr/local/lib/python3.9/site-packages/hailtop/batch/backend.py", line 595, in _run
    return async_to_blocking(
  File "/usr/local/lib/python3.9/site-packages/hailtop/utils/utils.py", line 156, in async_to_blocking
    return loop.run_until_complete(task)
  File "/usr/local/lib/python3.9/site-packages/nest_asyncio.py", line 81, in run_until_complete
    return f.result()
  File "/usr/local/Cellar/python@3.9/3.9.17_1/Frameworks/Python.framework/Versions/3.9/lib/python3.9/asyncio/futures.py", line 201, in result
    raise self._exception
  File "/usr/local/Cellar/python@3.9/3.9.17_1/Frameworks/Python.framework/Versions/3.9/lib/python3.9/asyncio/tasks.py", line 256, in __step
    result = coro.send(None)
  File "/usr/local/lib/python3.9/site-packages/hailtop/batch/backend.py", line 698, in _async_run
    used_remote_tmpdir_results = await bounded_gather(*[functools.partial(compile_job, j) for j in unsubmitted_jobs], parallelism=150)
  File "/usr/local/lib/python3.9/site-packages/hailtop/utils/utils.py", line 187, in bounded_gather
    return await gatherer.wait()
  File "/usr/local/lib/python3.9/site-packages/hailtop/utils/utils.py", line 248, in wait
    raise self._errors[0]
  File "/usr/local/lib/python3.9/site-packages/hailtop/utils/utils.py", line 224, in _worker
    res = await pf()
  File "/usr/local/lib/python3.9/site-packages/hailtop/batch/backend.py", line 695, in compile_job
    used_remote_tmpdir = await job._compile(local_tmpdir, batch_remote_tmpdir, dry_run=dry_run)
  File "/usr/local/lib/python3.9/site-packages/hailtop/batch/job.py", line 876, in _compile
    await self._batch._fs.write(code_path, job_command_bytes)
  File "/usr/local/lib/python3.9/site-packages/hailtop/aiotools/fs/fs.py", line 298, in write
    await retry_transient_errors(_write)
  File "/usr/local/lib/python3.9/site-packages/hailtop/utils/utils.py", line 813, in retry_transient_errors
    return await retry_transient_errors_with_debug_string('', 0, f, *args, **kwargs)
  File "/usr/local/lib/python3.9/site-packages/hailtop/utils/utils.py", line 825, in retry_transient_errors_with_debug_string
    return await f(*args, **kwargs)
  File "/usr/local/lib/python3.9/site-packages/hailtop/aiotools/fs/fs.py", line 296, in _write
    await f.write(data)
  File "/usr/local/lib/python3.9/site-packages/hailtop/aiotools/fs/stream.py", line 119, in __aexit__
    await self.wait_closed()
  File "/usr/local/lib/python3.9/site-packages/hailtop/aiotools/fs/stream.py", line 104, in wait_closed
    await self._wait_closed()
  File "/usr/local/lib/python3.9/site-packages/hailtop/aiocloud/aiogoogle/client/storage_client.py", line 85, in _wait_closed
    async with await self._request_task as resp:
  File "/usr/local/Cellar/python@3.9/3.9.17_1/Frameworks/Python.framework/Versions/3.9/lib/python3.9/asyncio/futures.py", line 287, in __await__
    return self.result()  # May raise too.
  File "/usr/local/Cellar/python@3.9/3.9.17_1/Frameworks/Python.framework/Versions/3.9/lib/python3.9/asyncio/futures.py", line 201, in result
    raise self._exception
  File "/usr/local/Cellar/python@3.9/3.9.17_1/Frameworks/Python.framework/Versions/3.9/lib/python3.9/asyncio/tasks.py", line 256, in __step
    result = coro.send(None)
  File "/usr/local/lib/python3.9/site-packages/hailtop/aiocloud/common/session.py", line 21, in post
    return await self.request('POST', url, **kwargs)
  File "/usr/local/lib/python3.9/site-packages/hailtop/aiocloud/common/session.py", line 105, in request
    return await self._http_session.request(method, url, **kwargs)
  File "/usr/local/lib/python3.9/site-packages/hailtop/httpx.py", line 137, in request_and_raise_for_status
    resp = await self.client_session._request(method, url, **kwargs)
  File "/usr/local/lib/python3.9/site-packages/aiohttp/client.py", line 535, in _request
    conn = await self._connector.connect(
  File "/usr/local/lib/python3.9/site-packages/aiohttp/connector.py", line 542, in connect
    proto = await self._create_connection(req, traces, timeout)
  File "/usr/local/lib/python3.9/site-packages/aiohttp/connector.py", line 907, in _create_connection
    _, proto = await self._create_direct_connection(req, traces, timeout)
  File "/usr/local/lib/python3.9/site-packages/aiohttp/connector.py", line 1206, in _create_direct_connection
    raise last_exc
  File "/usr/local/lib/python3.9/site-packages/aiohttp/connector.py", line 1175, in _create_direct_connection
    transp, proto = await self._wrap_create_connection(
  File "/usr/local/lib/python3.9/site-packages/aiohttp/connector.py", line 992, in _wrap_create_connection
    raise client_error(req.connection_key, exc) from exc
aiohttp.client_exceptions.ClientConnectorError: Cannot connect to host storage.googleapis.com:443 ssl:default [Too many open files]
bw2 commented 1 year ago

It seems to only occur when I use bp.read_input(..) to localize many files per job. Above a certain number of input files (many thousands), I started getting this error.

danking commented 12 months ago

Clarifying for future devs: the Batch.read_input is probably just causing the script to grow large enough that we start using scripts which need to be uploaded separately from the job.

Attempting to replicate with


In [6]: import hailtop.batch as hb
   ...: b = hb.Batch(backend=hb.ServiceBackend())
   ...: for _ in range(300):
   ...:     j = b.new_job()
   ...:     j.command(f'echo {"a" * 11 * 1024}')
   ...: b.run()

triggers https://github.com/hail-is/hail/issues/14051. I'll try to fix both.