Can't Create a Workspace using Google Cloud

[Hetnon]

Trying to create my first workspace. I hit create my space and I see 3 buckets being created on my GCP, but nothing shows up in the actual 'workspaces' in my databricks console. the only thing is the  'create workspace' button' also, there is no error message nowhere to be seen. I did it 3x and each time created 3 buckets and nothing else. I'm trying to test the LLM api to see if I switch from what I've been using now in my system. Is it available already? Seems the process is too clucky to simply using getting an API key to make an API call.

[Kaniz]

Hi @Hetnon, Creating a Databricks workspace can indeed be a bit tricky at times, but let’s troubleshoot this together!

Here are some steps you can take to address the issue:

1. Check Resource Provider Registration: Ensure that the Microsoft.Databricks resource provider is registered and has the status “Registered” in your Azure subscription. If it’s not registered, you can do so using the Azure CLI or PowerShell. Sometimes creating the workspace in a different Azure region might also help resolve the issue¹.

2. Storage Configuration:

   • If you encounter an error related to failed storage configuration validation checks, it might be due to incorrect S3 bucket permissions. Make sure your S3 bucket permissions are properly set up. You can follow the steps in the article on creating an S3 bucket for workspace deployment to ensure correctness.
   • Also, check if there are any issues related to the maximum number of VPCs, VPC endpoints, or Elastic IP Addresses. These limits might affect workspace creation, especially if you’re using a Databricks-managed VPC².

3. Credentials Configuration:

   • If you encounter credential configuration validation errors, pay attention to the specific permission checks mentioned in the error message.
   • If fewer than ten permission checks fail, your IAM policy might be missing specific permissions. You can copy the correct policy from the article on creating an IAM role for workspace deployment.
   • If ten or more checks fail, verify that the trust relationship of your customer role is correctly set up according to the instructions in the same article.
   • Confirm that you include the correct role ARN in the credentials object².

4. Update a Failed Workspace: If the status of your new workspace is “Failed”, you can update the failed workspace configuration and try creating it again with the new configuration. Log in to the account console as an account admin, navigate to the Workspaces page, click the name o...³.

If you encounter any further issues, feel free to ask for more assistance. Good luck with your Databricks journey! 😊

[Hetnon]

Hi, unfortunately I wasn't able to really grasp how to use the product. I had problems setting up the workspace and didn't understand how to setup and trial. I wanted to trial the LLM but couldn't.
I tried to deactivate the trial, which is coming to an end on the 22/04 but I'm not really sure if that worked too. The process was a whole really cumbersome.
I'm just writing to make sure I won't be billed anything at all related to the use of databricks. If you could reply just acknowledging this message and checking that the subscription was actually cancelled and no charge will incur until we properly understand the capabilities and can use it, I would much appreciate it.

And before ending, my message mentioned GCP which stands for Google Cloud Platform, and all your answers are related to AWS and their respective services. I would advise more careful answers to potential clients.

The email I'm sending this is not the main account user for the GCP, so I would appreciate a direct contact to check if the trial for the correct account has been terminated. I didn't find anywhere to contact Databricks directly.

Best Regards,
Hetnon Freitas