Welcome to the Microsoft Q&A forum! Thanks for your question.
If you are unable to create a Compute Instance in Azure Machine Learning Studio and the creation fails after several hours, please check the following:
Step 1: Check Compute Quotas & SKU Availability
- Ensure you have sufficient quota for the requested VM SKU.
- Some SKUs may not be available in your region.
- Refer to: 🔗 Azure Machine Learning Compute Instance Overview
Step 2: Known Issue: A10 SKU Not Supported
- The A10 GPU SKU is not supported for Azure ML Compute Instances.
- If you are trying to create an instance with this SKU, it will fail.
- More details here: 🔗 Known issues: A10 SKU Not Supported
Step 3: Create & Attach a Compute Cluster Instead
- If the instance is failing, you can try creating an Azure Machine Learning Compute Cluster instead.
- Step-by-step guide: 🔗 Create and Attach Compute Cluster
Step 4: Check Activity Logs & Retry
- Go to Azure Portal → Monitor → Activity Logs to check for failure details.
- Try using a different VM SKU or region.
Known Issue:
- Long Compute Instance Creation Time: Some users have reported that compute instance creation can take several hours due to region-specific limitations or VM SKU availability.
- Solution: Try selecting a different region, SKU, or use a Compute Cluster instead.
Please try these steps and let me know if you need further assistance, also request you to elaborate more steps of what you have did if you need more help so that I can reply back based on that. Hope this helps!
Regards,
Chakravarthi Rangarajan Bhargavi
Please kindly accept the answer and vote 'Yes' if you found it helpful to support the community. Thanks a lot!