Unable to see my artifacts on minio

[karthik kv]

Hi  All,

I have started using mlflow recently and I am having challenges get my models/artifacts in the minio which I am using as my object store.(I also don't see these artifacts in the mlflow server UI)

Details are as follows,

1. Using mlflow 2.2.0
2. python 3.11
3. Docker compose 2.2
4. SQLite as the backend metadata store
5. minio as the object store

when I run my pipeline I am able to see all the results in the mlflow server (running on 5000 port), including the logs and params, however they are not reflecting on the minio(both minio and mlflow are in the same network of a Dockercompose file).

Instead the artifacts are being created in a folder structure under

mlruns/<experiment_id>/<run_id>/artifacts/...

I am using the following command to the mlflow server  via my docker

mlflow server --backend-store-uri sqlite:////home/docker/data/database_files/mlops.db --default-artifact-root s3://tex-minio-bucket/mlflow -h 0.0.0.0 -p 5000

Note: I am using 9002 port on the laptop side and 9000 port on container side

1. Please suggest what should be the value of MLFLOW_S3_ENDPOINT_URL, Is it with port 9000 or 9002 and is with localhost or with the ip-address that I have defined in my docker compose?
2. What is correct command for the mlflow server when your bucket name is tex-minio-bucket and SQLite is the backend store
   

Any help is much appreciated and Waiting for your reply.

Regards,

Karthik