Conda run_env in custom docker image for Kuberay

I created a custom docker image with the following:

FROM rayproject/ray:2.10.0-py310


COPY conda.yml .

RUN conda env create -f conda.yml

Now when I run the following (note that I am kubectl port-forward to my cluster head service):

ray.init(address=“ray://localhost:10001/”, runtime_env={“conda”: “myenv”})

I get this:

Traceback (most recent call last):
File “/home/ray/anaconda3/lib/python3.10/site-packages/ray/util/client/server/”, line 719, in Datapath
raise RuntimeError(
RuntimeError: Proxy failed to Connect to backend! Check ray_client_server.err and ray_client_server_23009.err on the head node of the cluster for the relevant logs. By default these are located at /tmp/ray/session_latest/logs.

I do not get this error when running with runtime_env and can use the cluster as long as I dont utilize any dependencies in remote functions.

Not sure if I’m going about this the wrong way, but my assumption was that specifying the conda env name would activate for all future compute…