await file_obj.credential.close() : TypeError: object NoneType can't be used in 'await' expression

[ELToulemonde]

My problem

At the end of my python script, I get a clean up error : TypeError: object NoneType can't be used in 'await' expression

Complete trace is :

Traceback (most recent call last):
  File ".../lib/python3.10/weakref.py", line 667, in _exitfunc
    f()
  File ".../lib/python3.10/weakref.py", line 591, in __call__
    return info.func(*info.args, **(info.kwargs or {}))
  File ".../lib/python3.10/site-packages/fsspec/asyn.py", line 103, in sync
    raise return_result
  File ".../lib/python3.10/site-packages/fsspec/asyn.py", line 56, in _runner
    result[0] = await coro
  File ".../lib/python3.10/site-packages/adlfs/utils.py", line 78, in close_credential
    await file_obj.credential.close()
TypeError: object NoneType can't be used in 'await' expression

Which is weird because, I don't do anything asynchronious.

A reproducible example

At least as much as I can:

from azure.identity import ChainedTokenCredential, ManagedIdentityCredential, AzureCliCredential
azure_cli = AzureCliCredential()
managed_identity = ManagedIdentityCredential()
CREDENTIAL_CHAIN = ChainedTokenCredential(managed_identity, azure_cli)

import pandas as pd
pd.read_parquet("abfs://blob-name@datalake.blob.core.windows.net/path_to_parquets.parquet", storage_options={"credential": credential_chain})
print("Done")

I do get the "Done" printed before failure.

My config

• ubuntu,
• python 3.10,
• azure-storage-blob==12.16.0
• pandas==2.0.0
• pyarrow==11.0.0
• adlfs==2023.9.0

[TomAugspurger]

fsspec / adlfs use an async internally. can you try using the credentials from `azure.identity.aio` instead?

…

On Oct 6, 2023, at 10:43 AM, ELToulemonde ***@***.***> wrote: My problem At the end of my python script, I get a clean up error : TypeError: object NoneType can't be used in 'await' expression on Traceback (most recent call last): File ".../lib/python3.10/weakref.py", line 667, in _exitfunc f() File ".../lib/python3.10/weakref.py", line 591, in __call__ return info.func(*info.args, **(info.kwargs or {})) File ".../lib/python3.10/site-packages/fsspec/asyn.py", line 103, in sync raise return_result File ".../lib/python3.10/site-packages/fsspec/asyn.py", line 56, in _runner result[0] = await coro File ".../lib/python3.10/site-packages/adlfs/utils.py", line 78, in close_credential await file_obj.credential.close() TypeError: object NoneType can't be used in 'await' expression Which is weird because, I don't do anything asynchronious. A reproducible example At least as much as I can: from azure.identity import ChainedTokenCredential, ManagedIdentityCredential, AzureCliCredential azure_cli = AzureCliCredential() managed_identity = ManagedIdentityCredential() CREDENTIAL_CHAIN = ChainedTokenCredential(managed_identity, azure_cli) import pandas as pd ***@***.***/path_to_parquets.parquet", storage_options={"credential": credential_chain}) print("Done") I do get the "Done" printed before failure. My config ubuntu, python 3.10, azure-storage-blob==12.16.0 pandas==2.0.0 pyarrow==11.0.0 adlfs==2023.9.0 — Reply to this email directly, view it on GitHub <#431> or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAKAOIVBR7SYSQMVOV7SKMTX6ARJZBFKMF2HI4TJMJ2XIZLTSOBKK5TBNR2WLJDUOJ2WLJDOMFWWLO3UNBZGKYLEL5YGC4TUNFRWS4DBNZ2F6YLDORUXM2LUPGBKK5TBNR2WLJLJONZXKZNENZQW2ZNLORUHEZLBMRPXI6LQMWBKK5TBNR2WLJDUOJ2WLJDOMFWWLLTXMF2GG2C7MFRXI2LWNF2HTLDTOVRGUZLDORPXI6LQMWSUS43TOVS2M5DPOBUWG44SQKSHI6LQMWVHEZLQN5ZWS5DPOJ42K5TBNR2WLKJRGA4TIMBRHE4TTAVEOR4XAZNFNFZXG5LFUV3GC3DVMWVDCOJTGA2DOMBSGA2KO5DSNFTWOZLSUZRXEZLBORSQ>. You are receiving this email because you are subscribed to this thread. Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[davidsteinar]

@TomAugspurger @ELToulemonde I get the same error, no difference importing DefaultAzureCredential from either azure.identity.aio or azure.identity , did you solve this?

[mkp-jansen]

Same error here, any solutions?

[marktodisco]

Importing DefaultAzureCredential from azure.identity.aio silenced that error for me.

Python 3.10.13 on Ubuntu.

Package                     Version  
--------------------------- ---------
adlfs                       2023.12.0
azure-ai-ml                 1.12.1
azure-common                1.1.28
azure-core                  1.29.6
azure-datalake-store        0.0.53
azure-identity              1.15.0
azure-mgmt-core             1.4.0
azure-mgmt-resource         23.0.1
azure-mgmt-storage          21.1.0
azure-mgmt-subscription     3.1.1
azure-storage-blob          12.19.0
azure-storage-file-datalake 12.14.0
azure-storage-file-share    12.15.0
pyarrow                     14.0.2

[mhtrinh]

I am aware that using azure.identity.aio silence the error but: Why does async involved in non-async call ?

[TomAugspurger]

fsspec uses asyncio internally.

I'd recommend people use credentials from azure.identity.aio. If someone wants, we could add an inspect.iscoroutine check to before we call .close.

[mhtrinh]

That may solve one of our issue: I am using adlfs as part of a complex code that use ThreadPool. At the end of the run, I get this message that do not change the exit code, so not fatal but looks a bit ugly:

Traceback (most recent call last):
  File ".../lib/python3.10/weakref.py", line 667, in _exitfunc
    f()
  File ".../lib/python3.10/weakref.py", line 591, in __call__
    return info.func(*info.args, **(info.kwargs or {}))
  File ".../lib/python3.10/site-packages/fsspec/asyn.py", line 103, in sync
    raise return_result
  File ".../lib/python3.10/site-packages/fsspec/asyn.py", line 56, in _runner
    result[0] = await coro
  File ".../lib/python3.10/site-packages/adlfs/utils.py", line 78, in close_credential
    await file_obj.credential.close()
TypeError: object NoneType can't be used in 'await' expression

I did not manage to create a small reproducable example ...

[bikeshedder]

Reproduction example

Just create an AzureBlobFileSystem with non-AIO DefaultAzureCredential:

from adlfs import AzureBlobFileSystem
from azure.identity import DefaultAzureCredential

AzureBlobFileSystem(
    account_name="_",
    credential=DefaultAzureCredential(),
)

When the code exits an exception is raised:

Traceback (most recent call last):
  File "/usr/lib/python3.12/weakref.py", line 666, in _exitfunc
    f()
  File "/usr/lib/python3.12/weakref.py", line 590, in __call__
    return info.func(*info.args, **(info.kwargs or {}))
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File ".venv/lib/python3.12/site-packages/fsspec/asyn.py", line 103, in sync
    raise return_result
  File ".venv/lib/python3.12/site-packages/fsspec/asyn.py", line 56, in _runner
    result[0] = await coro
                ^^^^^^^^^^
  File ".venv/lib/python3.12/site-packages/adlfs/utils.py", line 78, in close_credential
    await file_obj.credential.close()
TypeError: object NoneType can't be used in 'await' expression

What about azure.identity.aio.DefaultAzureCredential

I'm using both azure.storage.blob.BlobServiceClient and adlfs.AzureBlobFileSystem in my code. I made the mistake importing the AIO version and using it for both which caused the ABFS to no longer raise an exception but BlobServiceClient stopped to work as it expects a sync version.

I fixed that by using non-AIO credentials for the BlobServiceClient and AIO ones for AzureBlobFileSystem. It's not pretty having to create two credential objects but fixes the issue.

[BnJam]

fsspec / adlfs use an async internally. can you try using the credentials from azure.identity.aio instead?

The azure.identity.aio solution worked for me in a simple script while testing out remote storage connectivity:

import io
import os
import adlfs

from azure.identity.aio import DefaultAzureCredential

credential = DefaultAzureCredential()

azfs = adlfs.AzureBlobFileSystem(
    account_name=<storage-account-name>,
    credential=credential
)

with io.BytesIO() as buf:
    buf.write(str('hello from the byterealm!').encode())
    buf.seek(0)
    azfs.write_bytes('/container/path/msg.txt', buf)

file_output = azfs.read_text('/container/path/msg.txt')

print(file_output)

Giving the print out:

$ python remote-storage.py 
hello from the byterealm!

That may solve one of our issue: I am using adlfs as part of a complex code that use ThreadPool. At the end of the run, I get this message that do not change the exit code, so not fatal but looks a bit ugly:

Traceback (most recent call last):
  File ".../lib/python3.10/weakref.py", line 667, in _exitfunc
    f()
  File ".../lib/python3.10/weakref.py", line 591, in __call__
    return info.func(*info.args, **(info.kwargs or {}))
  File ".../lib/python3.10/site-packages/fsspec/asyn.py", line 103, in sync
    raise return_result
  File ".../lib/python3.10/site-packages/fsspec/asyn.py", line 56, in _runner
    result[0] = await coro
  File ".../lib/python3.10/site-packages/adlfs/utils.py", line 78, in close_credential
    await file_obj.credential.close()
TypeError: object NoneType can't be used in 'await' expression

I did not manage to create a small reproducable example ...

Using an asynchronous object in a threadpool sounds like a headache... you could create a credential in each thread to isolate the async handlers per action, but, perhaps collecting the pool and finalising the credential once all threads are complete could work too? Practicing the former would help when you move from threads to processes or external Workers/Jobs.

[f-vt]

I had the same issue but I needed to use the exclude_interactive_browser_credential=False parameter which is not available in the azure.identity.aio one.

What I ended up is creating a custom class from the sync (not the aio), adding a dumb close async function so fsspec is happy:

class AsyncDefaultAzureCredentialWrapper(AsyncTokenCredential):
    """
    This class wraps the synchronous DefaultAzureCredential to provide an asynchronous interface.
    It allows the use of DefaultAzureCredential with fsspec, which requires asynchronous credentials.
    The close method is overridden to be a no-op to prevent issues when fsspec tries to close the credential.
    """

    def __init__(self, **kwargs):
        # Initialize the synchronous DefaultAzureCredential
        self._credential = DefaultAzureCredential(**kwargs)

    async def get_token(self, *scopes, **kwargs):
        #Asynchronously get a token for the specified scopes: this method runs the synchronous get_token method in an executor to avoid blocking.
        loop = asyncio.get_event_loop()
        token = await loop.run_in_executor(None, self._credential.get_token, *scopes, **kwargs)
        return AccessToken(token.token, token.expires_on)

    async def close(self):
        #No-op close method to prevent issues when fsspec tries to close the credential.
        pass

# Create the async credential from the sync one, with the no-op close
credential = AsyncDefaultAzureCredentialWrapper(exclude_interactive_browser_credential=False)

[mhtrinh]

Using an asynchronous object in a threadpool sounds like a headache... you could create a credential in each thread to isolate the async handlers per action, but, perhaps collecting the pool and finalising the credential once all threads are complete could work too? Practicing the former would help when you move from threads to processes or external Workers/Jobs.

The things is that I don't ever used async in my code nor import any aio ...
My import looks like this:

import os
from datetime import datetime

import fsspec
import s3fs
from urllib.parse import urlparse
from enum import Enum
from glob import glob
import shutil
import re
import json
import hashlib
import base64
import binascii
from adlfs import AzureBlobFileSystem
from azure.storage.blob import BlobServiceClient, ContentSettings
from azure.identity import DefaultAzureCredential, ManagedIdentityCredential
import dateutil

from iniconfig import IniConfig

from concurrent.futures import ThreadPoolExecutor, as_completed

Is there a flag to tell fsspec or adlfs to Not use async ??

[mhtrinh]

And if I use aio:

from azure.identity.aio import DefaultAzureCredential, ManagedIdentityCredential

I get:

raise Exception(future.exception())
Exception: 'coroutine' object has no attribute 'token'
sys:1: RuntimeWarning: coroutine 'DefaultAzureCredential.get_token_info' was never awaited

[nateprewitt]

This will likely be addressed with #537 once that gets merged.