DAB deploy missing packages

[Przemk00]

Hello everyone, not sure where to start but so will do it from the beginning. 

I encountered an issue where my .whl files in the dist/ and packages/ directories were not being deployed to Databricks using dab-deploy. After investigating, I discovered that these directories were listed in .gitignore, which prevented them from being included in the deployment. As a workaround, commenting out the dist/ and packages/ entries in .gitignore ensures the deployment is successful.

The question arises why for some it works and for others, it does not and has to do some workarounds.

Thanks,

P 

[-werners-]

it could be depending on the template you used when creating the package.
DAB is still rough around the edges.

[Przemk00]

Is there anything specific I should look in the template? 

[-werners-]

In your case it would be what is in the .gitignore file.
Somehow your files were ignored, and if you did not edit the .gitignore, it was probably copied from a template (or existing dab).

[saurabh18cs]

Production native approach is to maintain your artifacts in central antifactory and use this location in your .yml for libraries. This way you ensure your antifactory are properly versioned , governed and maintaining some sort of standards. I hope with this approach you can reduce your errors of missing packages/whl etc.

e.g.

        libraries:

        - pypi:

            package: requests==2.31.0

        - pypi:

            package: azure-eventhub==5.10.1

        - pypi:

            package: azure-identity==1.12.0

        - pypi:

            package: yourpackage==0.0.0

            repo: https://YOURPACKAGE_ARTIFACT_REPO:YOURPACKAGE_ARTIFACT_TOKEN@pkgs.dev.azure.com/asml/c3d5..............

[saurabh18cs]

Production native approach is to maintain your artifacts in central antifactory and use this location in your .yml for libraries. This way you ensure your antifactory are properly versioned , governed and maintaining some sort of standards. I hope with this approach you can reduce your errors of missing packages/whl etc.

e.g.

        libraries:

        - pypi:

            package: requests==2.31.0

        - pypi:

            package: yourpackage==0.0.0

            repo: https://YOURPACKAGE_ARTIFACT_REPO:YOURPACKAGE_ARTIFACT_TOKEN@pkgs.dev.azure.com/asml/c3d5..............