buntu/mlflow connection to mysql uses only default port

Bug Description

This issue tracks: https://bugs.launchpad.net/ubuntu-docker-images/+bug/2061031

I set up another container for mysql using port 33306 instead of the default 3306 because I had another mariaDB container already running on that port. I changed the docker compose entrypoint for mlflow to reflect the new port for mysql, but mlflow kept getting an error of connection refused. Nothing worked until I shut down the mariaDB container and reverted mysql to its default port of 3306. I either need help with the entrypoint to specify the correct port, or there is a bug in the mlflow container using the default port and ignoring the one in the entrypoint.

This did not work on the 33036 port:

    entrypoint: mlflow server --backend-store-uri mysql+pymysql://${DB_USER}:${DB_PASSWORD}@mysql_mlflow:33306/${DB_DATABASE} --default-artifact-root minio://${AWS_BUCKET_NAME}/ --artifacts-destination minio://${AWS_BUCKET_NAME}/ -h 0.0.0.0

Once I reset the mysql container to work off port 3306, the entrypoint worked:

    entrypoint: mlflow server --backend-store-uri mysql+pymysql://${DB_USER}:${DB_PASSWORD}@mysql_mlflow:3306/${DB_DATABASE} --default-artifact-root minio://${AWS_BUCKET_NAME}/ --artifacts-destination minio://${AWS_BUCKET_NAME}/ -h 0.0.0.0

This is the error I was getting from mlflow:

tracker_mlflow | 2024/04/11 13:52:37 WARNING mlflow.store.db.utils: SQLAlchemy engine could not be created. The following exception is caught.
tracker_mlflow | (pymysql.err.OperationalError) (2003, "Can't connect to MySQL server on 'mysql_mlflow' ([Errno 111] Connection refused)")
tracker_mlflow | (Background on this error at: https://sqlalche.me/e/14/e3q8)

To Reproduce

None

Environment

None

Relevant Log Output

Above

Additional Context

No response

canonical / base-mlflow