在 Python Azure Function 中使用 AzureML 時出現“失敗異常：OSError：[Errno 30] 只讀文件系統”

Question

問題

我正在嘗試准備，然后從 ZA78215F3532B5639 中的 Azure Function 向 Azure 機器學習提交一個新實驗。 因此，我使用dataset.register(...為我的 Azure ML 工作區注冊了一個新數據集，其中包含我的 ML model 的訓練數據。但是，當我嘗試使用以下代碼行創建此數據集時

dataset = Dataset.Tabular.from_delimited_files(path = datastore_paths)

然后我得到一個Failure Exception: OSError: [Errno 30] Read-only file system... 。

想法

1. 我知道如果可能的話，我不應該從 Azure function 中寫入文件系統。 但我實際上不想向本地文件系統寫入任何內容。 我只想在datastore_path下創建數據集作為對我的 blob 存儲的引用，然后將其注冊到我的 Azure 機器學習工作區。 但似乎from_delimited_files方法無論如何都試圖寫入文件系統（可能是一些緩存？）。
2. 我也知道有一個臨時文件夾，允許在其中寫入臨時文件。 但是，我相信我無法真正控制這種方法在哪里寫入數據。 我已經嘗試在使用os.chdir(tempfile.gettempdir())調用 function 之前將當前工作目錄更改為這個臨時文件夾，但這沒有幫助。

還有其他想法嗎？ 我不認為我在做一些特別不尋常的事情......

細節

我正在使用 python 3.7 和 azureml-sdk 1.9.0，我可以在本地毫無問題地運行 python 腳本。 我目前使用 Azure Functions 擴展版本 0.23.0（以及用於 CI/CD 的 Azure DevOps 管道）從 VSCode 部署。

這是我的完整堆棧跟蹤：

Microsoft.Azure.WebJobs.Host.FunctionInvocationException: Exception while executing function: Functions.HttpTrigger_Train
 ---> Microsoft.Azure.WebJobs.Script.Workers.Rpc.RpcException: Result: Failure
Exception: OSError: [Errno 30] Read-only file system: '/home/site/wwwroot/.python_packages/lib/site-packages/dotnetcore2/bin/deps.lock'
Stack:   File "/azure-functions-host/workers/python/3.7/LINUX/X64/azure_functions_worker/dispatcher.py", line 345, in _handle__invocation_request
    self.__run_sync_func, invocation_id, fi.func, args)
  File "/usr/local/lib/python3.7/concurrent/futures/thread.py", line 57, in run
    result = self.fn(*self.args, **self.kwargs)
  File "/azure-functions-host/workers/python/3.7/LINUX/X64/azure_functions_worker/dispatcher.py", line 480, in __run_sync_func
    return func(**params)
  File "/home/site/wwwroot/HttpTrigger_Train/__init__.py", line 11, in main
    train()
  File "/home/site/wwwroot/shared_code/train.py", line 70, in train
    dataset = Dataset.Tabular.from_delimited_files(path = datastore_paths)
  File "/home/site/wwwroot/.python_packages/lib/site-packages/azureml/data/_loggerfactory.py", line 126, in wrapper
    return func(*args, **kwargs)
  File "/home/site/wwwroot/.python_packages/lib/site-packages/azureml/data/dataset_factory.py", line 308, in from_delimited_files
    quoting=support_multi_line)
  File "/home/site/wwwroot/.python_packages/lib/site-packages/azureml/dataprep/api/readers.py", line 100, in read_csv
    df = Dataflow._path_to_get_files_block(path, archive_options)
  File "/home/site/wwwroot/.python_packages/lib/site-packages/azureml/dataprep/api/dataflow.py", line 2387, in _path_to_get_files_block
    return datastore_to_dataflow(path)
  File "/home/site/wwwroot/.python_packages/lib/site-packages/azureml/dataprep/api/_datastore_helper.py", line 41, in datastore_to_dataflow
    datastore, datastore_value = get_datastore_value(source)
  File "/home/site/wwwroot/.python_packages/lib/site-packages/azureml/dataprep/api/_datastore_helper.py", line 83, in get_datastore_value
    _set_auth_type(workspace)
  File "/home/site/wwwroot/.python_packages/lib/site-packages/azureml/dataprep/api/_datastore_helper.py", line 134, in _set_auth_type
    get_engine_api().set_aml_auth(SetAmlAuthMessageArgument(AuthType.SERVICEPRINCIPAL, json.dumps(auth)))
  File "/home/site/wwwroot/.python_packages/lib/site-packages/azureml/dataprep/api/engineapi/api.py", line 18, in get_engine_api
    _engine_api = EngineAPI()
  File "/home/site/wwwroot/.python_packages/lib/site-packages/azureml/dataprep/api/engineapi/api.py", line 55, in __init__
    self._message_channel = launch_engine()
  File "/home/site/wwwroot/.python_packages/lib/site-packages/azureml/dataprep/api/engineapi/engine.py", line 300, in launch_engine
    dependencies_path = runtime.ensure_dependencies()
  File "/home/site/wwwroot/.python_packages/lib/site-packages/dotnetcore2/runtime.py", line 141, in ensure_dependencies
    with _FileLock(deps_lock_path, raise_on_timeout=timeout_exception):
  File "/home/site/wwwroot/.python_packages/lib/site-packages/dotnetcore2/runtime.py", line 113, in __enter__
    self.acquire()
  File "/home/site/wwwroot/.python_packages/lib/site-packages/dotnetcore2/runtime.py", line 72, in acquire
    self.lockfile = os.open(self.lockfile_path, os.O_CREAT | os.O_EXCL | os.O_RDWR)

   at Microsoft.Azure.WebJobs.Script.Description.WorkerFunctionInvoker.InvokeCore(Object[] parameters, FunctionInvocationContext context) in /src/azure-functions-host/src/WebJobs.Script/Description/Workers/WorkerFunctionInvoker.cs:line 85
   at Microsoft.Azure.WebJobs.Script.Description.FunctionInvokerBase.Invoke(Object[] parameters) in /src/azure-functions-host/src/WebJobs.Script/Description/FunctionInvokerBase.cs:line 85
   at Microsoft.Azure.WebJobs.Script.Description.FunctionGenerator.Coerce[T](Task`1 src) in /src/azure-functions-host/src/WebJobs.Script/Description/FunctionGenerator.cs:line 225
   at Microsoft.Azure.WebJobs.Host.Executors.FunctionInvoker`2.InvokeAsync(Object instance, Object[] arguments) in C:\projects\azure-webjobs-sdk-rqm4t\src\Microsoft.Azure.WebJobs.Host\Executors\FunctionInvoker.cs:line 52
   at Microsoft.Azure.WebJobs.Host.Executors.FunctionExecutor.InvokeAsync(IFunctionInvoker invoker, ParameterHelper parameterHelper, CancellationTokenSource timeoutTokenSource, CancellationTokenSource functionCancellationTokenSource, Boolean throwOnTimeout, TimeSpan timerInterval, IFunctionInstance instance) in C:\projects\azure-webjobs-sdk-rqm4t\src\Microsoft.Azure.WebJobs.Host\Executors\FunctionExecutor.cs:line 587
   at Microsoft.Azure.WebJobs.Host.Executors.FunctionExecutor.ExecuteWithWatchersAsync(IFunctionInstanceEx instance, ParameterHelper parameterHelper, ILogger logger, CancellationTokenSource functionCancellationTokenSource) in C:\projects\azure-webjobs-sdk-rqm4t\src\Microsoft.Azure.WebJobs.Host\Executors\FunctionExecutor.cs:line 532
   at Microsoft.Azure.WebJobs.Host.Executors.FunctionExecutor.ExecuteWithLoggingAsync(IFunctionInstanceEx instance, ParameterHelper parameterHelper, IFunctionOutputDefinition outputDefinition, ILogger logger, CancellationTokenSource functionCancellationTokenSource) in C:\projects\azure-webjobs-sdk-rqm4t\src\Microsoft.Azure.WebJobs.Host\Executors\FunctionExecutor.cs:line 470
   at Microsoft.Azure.WebJobs.Host.Executors.FunctionExecutor.ExecuteWithLoggingAsync(IFunctionInstanceEx instance, FunctionStartedMessage message, FunctionInstanceLogEntry instanceLogEntry, ParameterHelper parameterHelper, ILogger logger, CancellationToken cancellationToken) in C:\projects\azure-webjobs-sdk-rqm4t\src\Microsoft.Azure.WebJobs.Host\Executors\FunctionExecutor.cs:line 278
   --- End of inner exception stack trace ---
   at Microsoft.Azure.WebJobs.Host.Executors.FunctionExecutor.ExecuteWithLoggingAsync(IFunctionInstanceEx instance, FunctionStartedMessage message, FunctionInstanceLogEntry instanceLogEntry, ParameterHelper parameterHelper, ILogger logger, CancellationToken cancellationToken) in C:\projects\azure-webjobs-sdk-rqm4t\src\Microsoft.Azure.WebJobs.Host\Executors\FunctionExecutor.cs:line 325
   at Microsoft.Azure.WebJobs.Host.Executors.FunctionExecutor.TryExecuteAsyncCore(IFunctionInstanceEx functionInstance, CancellationToken cancellationToken) in C:\projects\azure-webjobs-sdk-rqm4t\src\Microsoft.Azure.WebJobs.Host\Executors\FunctionExecutor.cs:line 117

Answer 1

問題是我的虛擬環境中的操作系統版本不兼容。

非常感謝PramodValavala-MSFT提出創建 docker 容器的想法，根據他的建議，我突然收到以下有關dataset = Dataset.Tabular.from_delimited_files(path = datastore_paths)命令的錯誤消息：

異常：NotImplementedError：不支持 Linux 分布 debian 10。

這讓我想起了 azure 機器學習文檔中的以下警告：

某些數據集類依賴於 azureml-dataprep package，它僅與 64 位 Python 兼容。 For Linux users, these classes are supported only on the following distributions: Red Hat Enterprise Linux (7, 8), Ubuntu (14.04, 16.04, 18.04), Fedora (27, 28), Debian (8, 9), and CentOS ( 7）。

Choosing the predefined docker image 2.0-python3.7 (running Debian 9) instead of 3.0-python3.7 (running Debian 10) solved the issue (see https://hub.docker.com/_/microsoft-azure-functions-蟒蛇）。

我懷疑我最初使用的默認虛擬環境也在不兼容的操作系統上運行。