Flyte enables production-grade orchestration for machine learning workflows and data processing created to accelerate local workflows to production.

Flyte

Hi, does someone has some info about the correct setup of env vars/configuration file for that? I’d like to avoid to have going to investigate inside the source code

<@U07BEHZQU9X> so it seems it's effectively trying to register but failing to contact the `minio` storage layer, this is the NodePort that exposes the `30002` your pipeline is failing to access

```flyte-sandbox-minio       9000:30002/TCP,9001:31875/TCP   44d ```

configuration for minio might be more involved:
<https://github.com/davidmirror-ops/flyte-the-hard-way/blob/7c91dae9fbd725cb4efb2628f678b4d4a0024e4b/docs/on-premises/single-node/manifests/onprem-flyte-binary-values.yaml#L6-L29|https://github.com/davidmirror-ops/flyte-the-hard-way/blob/7c91dae9fbd725cb4efb262[…]-premises/single-node/manifests/onprem-flyte-binary-values.yaml>

Hi <@U04H6UUE78B>,  yes the problem is that I am running `pyflyte register` from a kubernetes job running inside the Flyte demo cluster and I don’t find a way to make it point to the right minio url, because it always uses `localhost, 30002` setting. I tried to set the `FLYTE_AWS_ENDPOINT: "<http://minio.flyte.svc.cluster.local:9000>"` but on my job and on `flyte-sandbox` pod but nothing.
My flyte demo cluster settings are the default ones.

I got it to run, I had to comment out the rows below in the `flyte-sandbox-config` configmap
```    # storage:
    #   signedURL:
    #     stowConfigOverride:
    #       endpoint: <http://localhost:30002>```
I guess you use it somewhere in the frontend to redirect the user to the minio FE

Thank you for pointing out the file, I got from it to go looking at the configmaps mounted on the flyte pod

ah yes, this is used by the client but since you're running inside the cluster, safe to ignore it