Mlflow with ray core

1

mlflow run is not resumed in same run_id with parallel iterators
PS: example

import ray

ray.init(address="auto")
from mlflow import log_metric
import mlflow





@ray.remote
def task(client_id,run_id):
    val = client_id + 10
    with mlflow.start_run(run_id=run_id, nested=True):
        log_metric(f"global_accuracy", val, step=1)


if __name__ == "__main__":
    mlflow.set_tracking_uri("sqlite:///ml.db")
    mlflow.start_run()
    run_id = mlflow.active_run().info.run_id
    ray.get([task.remote(i,run_id) for i in range(6)])
2

i get the following error

store.py", line 549, in _get_run_info
raise MlflowException(
mlflow.exceptions.MlflowException: Run ‘12a3b8f2aba1439794210b5b9e5bb854’ not found

3

Could you actually ask in the ray core category?