Hello, If you install Apache Airflow 3.3.0 with the official constraints on Python 3.11 or 3.12, the databricks provider was being pinned to an old release (apache-airflow-providers-databricks 6.13.0, from October 2024), which caused import errors. This happened because thrift 0.23.0 conflicted with databricks-sql-connector 4.x during dependency resolution, downgrading the connector and dragging the provider back to the old version.
This is now fixed: - The constraints-3.3.0 tag and the constraints-3-3 branch now pin apache-airflow-providers-databricks 7.16.1, databricks-sql-connector 4.2.5 and thrift 0.16.0 for Python 3.11 and 3.12 (matching the other interpreters). - The official apache/airflow:3.3.0 Python 3.11 and 3.12 Docker images have been rebuilt with the corrected constraints. No action is needed beyond re-pulling the 3.3.0 image or reinstalling with the updated constraints. Details: https://github.com/apache/airflow/issues/69603 Docs: https://airflow.apache.org/docs/docker-stack/changelog.html#changes-after-publishing-the-images Regards, Jarek --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
