How to fix "The pipeline compute target mycluster is invalid." in Azure ML Designer?

Question

Hi everyone,

I am following Microsoft Learn instructions at <<https://microsoftlearning.github.io/AI-900-AIFundamentals/instructions/02c-create-clustering-model.html>>. However, I am facing an issue in Azure Machine Learning Designer.

I created a compute cluster named "mycluster", Compute type: Machine Learning compute, size: Standard_DS11_v2 (2 cores, 14 GB RAM, 28 GB disk), Processing unit: "CPU - Memory optimized", OS type: Linux, Virtual machine tier: dedicated, Current available nodes: 2, Minimum number of nodes: 0, Maximum number of nodes: 2, Idle seconds before scale down: 120, SSH access: disabled, Region: brazilsouth (same region as my Azure ML workspace named "myworkspace"), State: Succeeded (0 nodes), Active runs: 0, Idle nodes: 0, Busy nodes: 0, Unprovisioned nodes: 2, Workspace: myworkspace. Both myworkspace and mycluster are in the same resource group.

Then I created and saved a pipeline following Microsoft Learn directions. In order to apply the data transformations, I need to run the pipeline as an experiment, as observed by the Microsoft Learn documentation provided in the link.

1. I select Configure & Submit at the top of the page to open the Set up pipeline job dialogue
2. On the Basics page I select Create New and set the experiment name
3. On Inputs & outputs I still follow MS Learn instructions and click Next
4. On the Runtime settings in the Select Compute Type I select Compute Cluster, and in the Select Azure ML Compute Cluster I select the recently create compute cluster "mycluster". That is when Azure shows the following error message under "Default compute": "The pipeline compute target mycluster is invalid."

I tried to create compute clusters with different sizes and different numbers of nodes, and I even tried to create a compute instance with same size and specs as the compute cluster but all of the alternatives generate the same error message.

I am including the Q&A Assist's answer. As for suggested step 1, I do not see any "fails" when I select mycluster, and look at the cluster info that (as provided on the top of this post). I believe the compute target is properly set up and configured. The suggested step 2 is somehow like the step 1 but since I do not see any differences, this could be the cause of the error. Finally, the step 3 is, I believe, what I am (unsuccessfully) trying to do.

I strongly appreciate any help provided. Sincerely,

Marco

Answer 1

I found a solution for the issue. I just deleted my existing compute cluster (size Standard_DS11_v2) e created a compute instance using a more powerfull specification (size Standard_E4ds_v4), and now I can use the Azure Machine Learning Designer! Hope that helps anyone with same issue. Best, [m]

Answer 2

I found a solution for the issue. I just deleted my existing compute cluster (size Standard_DS11_v2) e created a compute instance using a more powerfull specification (size Standard_E4ds_v4), and now I can use the Azure Machine Learning Designer! Hope that helps anyone with same issue. Best, [m]

Answer 3

The error message "The pipeline compute target mycluster is invalid" in Azure Machine Learning Designer may occur due to various reasons. One possible reason is that the compute target is not properly set up or configured. To fix this issue, you can try the following steps:

1. Check if the compute target is properly set up and configured. Make sure that the compute target is in the same region as your workspace, and that it is in a running state. Also, make sure that the compute target is of the correct type (e.g., Machine Learning Compute) and that it has the required resources (e.g., CPU, memory, disk).
2. Check if the compute target is properly authorized. Make sure that the compute target is authorized to access your workspace, and that it has the required permissions to run the pipeline. You can check the authorization status of the compute target in the Azure portal or using the Azure Machine Learning SDK.
3. Check if the pipeline is properly configured. Make sure that the pipeline is properly configured to use the compute target. You can check the pipeline configuration in the Azure Machine Learning Designer or using the Azure Machine Learning SDK.

If none of the above steps work, you can try deleting the compute target and creating a new one. Also, make sure that you are following the latest instructions and best practices for using Azure Machine Learning Designer.

---

References:

• Example pipelines & datasets for Azure Machine Learning designer
• Rebuild a Studio (classic) experiment in Azure Machine Learning
• Create and run machine learning pipelines using components with the Azure Machine Learning studio