SSL error on ws.compute_targets

[AndrewLouw]

Trying to get ws.compute_targets hangs for a long time (max retries) then eventually produces an error. The actual error is:
Max retries exceeded
but this is due to an underlying bad handshake

This does not happen on 1.0.85

To replicate:

1. create a new azureml workspace resource
2. pip install --upgrade azure-sdk
3. download config file
   4 run below:

import azureml
from azureml.core import Workspace
print(f"version: {azureml.core.version}")
ws = Workspace.from_config(config)
ws.compute_targets

Out:
version: 1.12.0

...Traceback...

SSLError: HTTPSConnectionPool( REDACTED_PII ) Max retries exceeded with url: REDACTED_PII api-version=2020-02-02 (Caused by SSLError(SSLError("bad handshake: Error([('SSL routines', 'tls_process_server_certificate', 'certificate verify failed')],)",),))

---

Document Details

⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.

• ID: 90efe042-27ac-6e77-9292-10df77bb1c36
• Version Independent ID: 3b9f2576-7d3f-00f8-0afe-549960520f08
• Content: azureml.core.Workspace class - Azure Machine Learning Python
• Content Source: AzureML-Docset/stable/docs-ref-autogen/azureml-core/azureml.core.Workspace(class).yml
• Service: machine-learning
• Sub-service: core
• GitHub Login: @DebFro
• Microsoft Alias: debfro

[nishankgu]

#assign @sauryadas

[AndrewLouw]

I got this issue on release day for version: 1.12.0 (and I suspect new infrastructure on cloud service).

[sauryadas]

Can you please provide information on which region you deployed? is this AKS ? Did you create or attach the cluster? are you able to access the cluster with the kube config?

[AndrewLouw]

I was using an Azure ML workspace in the UK South region.
Here is a corrected and improved set of reproduction steps - I tested today and still got the error:

1. create a new azureml workspace resource: https://portal.azure.com/#create/Microsoft.MachineLearningServices
2. pip install --upgrade azureml-mlflow
3. pip install --upgrade azureml-sdk
4. download config file
5. run below:

import azureml
from azureml.core import Workspace
print(f"version: {azureml.core.__version__}")
ws = Workspace.from_config()
ws.compute_targets

[sauryadas]

Do you have a correlation id?

Very hard to look at this and pin point the error. Are you able to create a support ticket?

[aashishb]

Is it OK to close this issue if it is being handled via support ticket?

[lostmygithubaccount]

no

[shayeny]

Was this solved? I am facing the same issue trying to pip install from behind a VNET.

[sauryadas]

I notice that this is issue was in v 1.12.0

Can you try the latest SDK version ? v1.22.0

https://docs.microsoft.com/en-us/azure/machine-learning/azure-machine-learning-release-notes?view=azure-ml-py&toc=https%3A%2F%2Fdocs.microsoft.com%2Fen-us%2Fpython%2Fapi%2Foverview%2Fazure%2Fml%2Ftoc.json%3Fview%3Dazure-ml-py&bc=https%3A%2F%2Fdocs.microsoft.com%2Fen-us%2Fpython%2Fazureml_stable_breadcrumb%2Ftoc.json%3Fview%3Dazure-ml-py

[v-strudm-msft]

We haven't received a response to our last comment so assume it's resolved and are closing it. Should there still be a question, please reopen this issue. Thank you.