`InvalidResourceName` when trying to deploy a model via the CLI to an existing resource

Question

Following this module: https://learn.microsoft.com/en-us/training/modules/get-started-openai/5-deploy-models

Note, this works fine in Azure OpenAI Studio. I only get errors when deploying using the CLI.

I'm running this command (not using my real resource name, but mine is also all lowercase letters, no symbols, numbers, etc.):

az cognitiveservices account deployment create -g my-openai-rg -n myresourcename --model-name gpt-35-turbo --model-version "0613" --model-format OpenAI --sku-name "Standard" --sku-capacity 1

The error is:

(InvalidResourceName) The resource name is invalid. Resource name can only include alphanumeric characters, underscores and hyphens; Resource name only allows 2 to 64 characters.

Code: InvalidResourceName

Message: The resource name is invalid. Resource name can only include alphanumeric characters, underscores and hyphens; Resource name only allows 2 to 64 characters.

The resource name is correct. When I tried using a resource name that doesn't exist, I get a different error.

Is this a bug?

Answer 1

Hi Tatiana Racheva,
Since the Microsoft Q&A community has a policy that the question author cannot accept their own answer. They can only accept answers by others, Summarized and posted the solution that you shared, request you to Accept the answer, it will benefit the community find the answers quickly.

Scenario: InvalidResourceName error when trying to deploy a model via the CLI to an existing resource.

Issue: The deployment command fails with an InvalidResourceName error, despite the resource name being correct.

Things Tried: Running the command without the --deployment-name parameter.

Solution shared by @Tatiana Racheva

Resolution: This worked: Adding the --deployment-name parameter to the command. This parameter is required to specify the name of the deployment.

Answer 2

Of course, as soon as I post it, I figure out the problem.

The error message is wrong (again, nothing wrong with the resource name), but the error is legitimate. I was missing the --deployment-name parameter.

Once I added it, the deployment was created.