[BUG] AzureML on Azure Functions Consumption Plan Linux

Investigative information

Please provide the following:

• Function App name: happy to provide privately
• Core Tools version: V4

Repro steps

Using the azureml package in Python in an Azure Function on a consumption plan (linux) breaks for basic operations such as

Dataset.to_pandas_dataframe()

with the following error message:

Exception: NotImplementedError: Linux distribution debian 11. does not have automatic support.
Missing packages: {'liblttng-ust.so.0'}
.NET Core 3.1 can still be used via `dotnetcore2` if the required dependencies are installed.
Visit https://aka.ms/dotnet-install-linux for Linux distro specific .NET Core install instructions.
Follow your distro specific instructions to install `dotnet-runtime-*` and replace `*` with `3.1.23`.

Expected behavior

I expect a Microsoft stack such as Azure Functions + azureml + .NET to be fully integrated and to “just work”.

My requirements as they relate to Azure packages:

azure-functions==1.9.0
azure-functions-durable==1.1.3
azure-storage-blob==12.9.0
azureml-core==1.43.0
azureml-dataset-runtime==1.43.0

The same bug has already been raised and unfortunately closed (#1007) but the solution proposed therein doesn’t seem helpful: I’m trying to use a tech stack that is fully MS / Azure (Azure Functions + azureml) so I don’t expect the solution here is to build my own Docker image. I would expect that the base Docker image underlying Azure Functions consumption plan includes the MS technology .NET if it’s necessary.

[ramya894]

@waltherg We will discuss this issue with our next level team and update you.

[pdthummar]

hi @waltherg, we are trying to reproduce the issue and investing it. Unfortunately, at this time, we cannot provide any timelines without completely investing the issue. Thanks