topic Re: DLT error at validation in Data Engineering

DLT error at validation

AlvaroCM — Wed, 11 Sep 2024 09:15:08 GMT

Hello,

I'm creating a DLT pipeline with Databricks on AWS. After creating an external location for my bucket, I encountered the following error:

DataPlaneException: [DLT ERROR CODE: CLUSTER_LAUNCH_FAILURE.CLIENT_ERROR] Failed to launch pipeline cluster aaaaa-bbbbb-cccc with termination code AWS_AUTHORIZATION_FAILURE and termination type CLIENT_ERROR: Could not launch cluster due to cloud provider authorization failure. databricks_error_message: Failure happened when talking to AWS, AWS API error code: AccessDenied AWS error message: User: arn:a...

This is the full error that the DLT pipeline shows, even in JSON format, so I'm unsure which AWS ARN it's referring to.

Steps to reproduce the error:

I created an External Location.
I attached an empty Notebook script to my Delta Live Tables pipeline (I just want to validate it so far).

When I press "Validate," it throws that error at the "waiting for resources" step in the DLT graph.

My guess is that I'm missing some additional cluster authorization in AWS (IAM) for DLT that I couldn't find in the Databricks documentation.

Thanks!

Re: DLT error at validation

MuthuLakshmi — Thu, 07 Nov 2024 16:28:56 GMT

@AlvaroCM
The issue could be because the AWS doesn't have the instance type "waiting for resources"

Can you use the instance type of the cluster which you were able to create using the clusters UI explicitly over the DLT pipeline as mentioned below.

"node_type_id": "m4.4xlarge",
"driver_node_type_id": "m4.4xlarge"

Re: DLT error at validation

AlvaroCM — Fri, 08 Nov 2024 09:02:31 GMT

Hi!

The error was related to the roles and permissions created when the workspace was set up. I reloaded the setup script in a new workspace, and it worked without problems.

Hope it helps anyone in the future.

Thanks!