Virtual Machine with Ubuntu Data Science Image and Standard NV18ads A10 v5 fails to boot

Question

I am trying to create a virtual machine using the data science image for ubuntu together with Standard NV18ads A10 v5 GPU. However, I get the following error message on each start/restart:

(OSProvisioningTimedOut) OS provisioning failure has reached terminal state and is non-recoverable for VM 'x'. Consider deleting and recreating this virtual machine. Additional Details: OS Provisioning for VM 'x' did not finish in the allotted time. The VM may still finish provisioning successfully. Please check provisioning state later. For details on how to check current provisioning state of Windows VMs, refer to https://aka.ms/WindowsVMLifecycle and Linux VMs, refer to https://aka.ms/LinuxVMLifecycle.

I tried recreating this VM, but the problem is still present. How can i resolve this issue?

Answer 1

To resolve the OSProvisioningTimedOut error when creating an Azure VM:

1. Check Compatibility

Verify that the selected Standard NV18ads A10 v5 GPU size supports the chosen image (e.g., Data Science Ubuntu). Use a compatible image or Ubuntu Server.

2. Try Another Image

If issues persist, use Ubuntu Server and install required tools manually.

3. Enable Boot Diagnostics

Enable boot diagnostics in the Azure Portal to check VM startup logs for errors.

4. Verify Network Settings

Ensure outbound internet access is allowed via the NSG and no restrictive firewall rules are blocking provisioning.

5. Recreate VM

Delete the VM and recreate it using Azure CLI or Portal:

az vm create \

  --resource-group MyResourceGroup \

  --name MyVM \

  --image microsoft-dsvm:ubuntu-data-science-vm:ubuntu-20-04:latest \

  --size Standard_NV18ads_A10_v5

6. Update Drivers

SSH into the VM after creation and install the NVIDIA driver:

sudo apt update

sudo apt install nvidia-driver-510

7. Contact Support

If the issue persists, contact Azure Support with boot diagnostics logs.

Answer 2

To resolve the OSProvisioningTimedOut error when creating an Azure VM:

1. Check Compatibility

Verify that the selected Standard NV18ads A10 v5 GPU size supports the chosen image (e.g., Data Science Ubuntu). Use a compatible image or Ubuntu Server.

2. Try Another Image

If issues persist, use Ubuntu Server and install required tools manually.

3. Enable Boot Diagnostics

Enable boot diagnostics in the Azure Portal to check VM startup logs for errors.

4. Verify Network Settings

Ensure outbound internet access is allowed via the NSG and no restrictive firewall rules are blocking provisioning.

5. Recreate VM

Delete the VM and recreate it using Azure CLI or Portal:

az vm create \

  --resource-group MyResourceGroup \

  --name MyVM \

  --image microsoft-dsvm:ubuntu-data-science-vm:ubuntu-20-04:latest \

  --size Standard_NV18ads_A10_v5

6. Update Drivers

SSH into the VM after creation and install the NVIDIA driver:

sudo apt update

sudo apt install nvidia-driver-510

7. Contact Support

If the issue persists, contact Azure Support with boot diagnostics logs.

Answer 3

Hi Fardokht,

Welcome to the Microsoft Q&A Platform! Thank you for asking your question here.

Thank you, M A Nakib, for sharing your experience. Here are additional steps to try:

The OS provisioning failure you're encountering can be due to several reasons. First, make sure that the Standard NV18ads A10 v5 GPU is compatible with the Ubuntu Data Science image you're using. Sometimes, provisioning failures are caused by insufficient resources in the selected region, so check if there are enough GPU resources available. You can also enable boot diagnostics to check the VM logs for more detailed error information. If the issue persists, consider using a different Ubuntu image or recreate the VM with different settings. You can verify supported sizes for GPU-based images in Azure by visiting the Azure VM Size documentation. Also, refer to the Linux VM Lifecycle documentation for troubleshooting steps. If none of these solutions work, contacting Azure support would be the next best step.

If an answer has been helpful, please consider accept the "Answer" and "Upvote" to help increase visibility of this question for other members of the Microsoft Q&A community.