cancel
Showing results for 
Search instead for 
Did you mean: 
Data Engineering
Join discussions on data engineering best practices, architectures, and optimization strategies within the Databricks Community. Exchange insights and solutions with fellow data engineers.
cancel
Showing results for 
Search instead for 
Did you mean: 

Redeploy Databricks Asset Bundle created by others

echol
Visitor

Hi everyone,

Our team is using Databricks Asset Bundles (DAB) with a customized template to develop data pipelines. We have a core team that maintains the shared infrastructure and templates, and multiple product teams that use this template to develop and deploy their own pipelines.

One common use case is the following:

  • Staff A (from Team AA) develops a pipeline using the DAB template and deploys the bundle to the QA environment for testing.

  • Later, Staff B (from the core team) needs to apply some changes to the same bundle (for example, updating job configuration or infrastructure-related settings).

However, the job created by the bundle is owned by Staff A. When Staff B tries to redeploy the same bundle with modifications, they receive an error like:

Error: permission denied creating or updating job_XXXXX. For assistance, contact the owners of this project.

This appears to be a job permission issue. We have tried configuring a service principal for “run as”, but the error still occurs.

Since the core team needs to make this type of change frequently, it’s not practical for us to constantly coordinate with the original job owner (Staff A) to redeploy bundles.

Is it possible for one user to redeploy or update a DAB that was initially deployed by another user?
If so, what is the recommended pattern or best practice for handling this in a multi-team setup?

Thanks!

1 REPLY 1

pradeep_singh
New Contributor II

Assuming the workflow is being deployed with dabs in production mode ,you can assign CAN_MANAGE permission on the workflow to a user group . anyone in this group can now deploy the workflow with dabs 

In higher environments you should be using a service principal and production mode to deploy your workflow and not individual users