Yuki Watanabe

The error message improvement mentioned [here](https://github.com/mlflow/mlflow/issues/12219#issuecomment-2154936557) is done by https://github.com/mlflow/mlflow/commit/726817118ae713674a33707bf6c9f43e168a4d48. The other issue mentioned above is being resolved in https://github.com/mlflow/mlflow/pull/13566.

Thank you for reporting the issue, @julcsii !! I could reproduce the issue, and it seems the error happens when getting first 5 rows for computing dataset digest [here](https://github.com/mlflow/mlflow/blob/3d9d67f596e03e868d79d6a27b48cec4882839da/mlflow/models/evaluation/base.py#L579). If...

@serena-ruan > I feel like the nested structure is because originally the arguments passed are nested This is True, but this doesn't necessarily mean we need nested context managers. Two...

Resolved in https://github.com/mlflow/mlflow/commit/726817118ae713674a33707bf6c9f43e168a4d48.

Hmm, while `log_datasets` is no-op, it seems there are a few places we pass the parameter; - https://github.com/mlflow/mlflow/blob/924e614462b76fc059aa55fe3ab98092c03815f0/mlflow/tracking/fluent.py#L2293 - https://github.com/mlflow/mlflow/blob/924e614462b76fc059aa55fe3ab98092c03815f0/mlflow/utils/autologging_utils/__init__.py#L68 Especially for the first one, it seems we need to...

@thatcort Thank you for reporting the issue! There is an ongoing effort for supporting Pyarrow 16: https://github.com/mlflow/mlflow/pull/11905.

@rysweet Could you kindly provide an update on this?

@huankimtran This will be available in a patch release next week.

Pyarrow pin was updated to newer than 16 in MLflow 2.16.0. The latest MLflow version support pyarrow up to 21.

@zmunro Could you share a bit more about your environment set up, particularly the type of backend store? One potential cause I suspect is you are not setting `--backend-store-uri` when...