Azure-docs: Creating an environment locally, or in Azure?

[DaveVoyles]

"To deploy and test your web service on the local machine, set up a local environment using the following command. The resource group name is optional."

az ml env setup -l [Azure Region, e.g. eastus2] -n [your environment name] [-g [existing resource group]]

This looks like it is creating an environment in Azure, and not locally. Am I misunderstanding?

---

Document Details

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

• ID: 5904637d-335b-ed1e-2f81-646a3a83a80c
• Version Independent ID: 91bd13ab-e31f-9bd9-f869-fa1e9e03000a
• Content: Azure Machine Learning Model Management Setup and Configuration | Microsoft Docs
• Content Source: articles/machine-learning/preview/deployment-setup-configuration.md
• Service: machine-learning
• GitHub Login: @raymondlaghaeian
• Microsoft Alias: raymondl

[jason-j-MSFT]

@DaveVoyles
Thanks for the feedback! We are currently investigating and will update you shortly.

[mimckitt]

@DaveVoyles It uses a combination of both. If you look at the difference between a Local Deployment and a Cluster deployment the only thing you will notice is a cluster deployment creates a Kubernetes Deployment on ACS.

So essentially you are using your local machine as the Container Service for the deployment.

Once you confirm it can run locally you can perform the Cluster Deployment which takes your local machine out of the mix. Does that make sense?

The local environment setup command creates the following resources in your subscription:
-A resource group (if not provided, or if the name provided does not exist)
-A storage account
-An Azure Container Registry (ACR)
-An Application insights account

The cluster environment setup command creates the following resources in your subscription:
-A resource group (if not provided, or if the name provided does not exist)
-A storage account
-An Azure Container Registry (ACR)
-A Kubernetes deployment on an Azure Container Service (ACS) cluster
-An Application insights account

[DaveVoyles]

OK, I think I have a better understanding now, thank you.

So if I were to go with the first option (local), it will prepare the setup for a docker container to run locally (in this case, OS X), as well as prep the environment in Azure. For a single container.

And I suppose the idea is that when this gets deploy to Azure, it will actually be deployed in a Docker container to a VM.

Thank you for clarifying. I ask these questions because we have an upcoming workshop with customers in LA next week, and I just wanted to verify that I was understanding everything correctly. My alias is Dvoyles if it's easier to speak there as well.

[mimckitt]

Happy to help! What you said is correct. Generally you will run something locally to verify that the code works then proceed with pushing it to Azure. Overall this saves you time as you don't need to constantly redeploy to Azure to test simple code changes. Especially if you are like me and constantly forgot ";" and quotes :)

[DaveVoyles]

Perfect, thank you! Actually just got my first web deploy working while you write were thing that, so it all went well!

[mimckitt]

@DaveVoyles Awesome! Best of luck in your upcoming workshop. I will go ahead and close this out but if you feel more clarification is needed just let me know.