Hello, I have to configure my local flyte config to access minio under a specific port. I assumed that would happen in my

.flyte/config.yaml

.

admin:
  # For GRPC endpoints you might want to use dns:///flyte.myexample.com
  endpoint: dns:///<my-flyte-domain>
  authType: Pkce
storage:
  connection:
    endpoint: http://<mycustomminiodomain>:<custom-port>

Wouldn't I do that over the

endpoint

setting? After adjusting the config

pyflyte run --remote

still runs again

host='minio.flyte.svc.cluster.local', port=9000

Hey Broder So this it set at the platform level but you need to override it? Or it would be better to set the custom port at the platform level?

Do you mean by platform level the environment where flyte is deployed in?

I mean, at the Helm chart config level

at helm level I've setup minio with default parameters I think. Services looks like this:

flyte                  minio                                NodePort       <someip>    <none>        9000:30084/TCP,9001:30088/TCP

My issue is that I won't get port 9000 opened on the on premise server soon and therefore thought I could just utilize the open port

30084

as well for my

pyflyte

execution

so, in that case you should be port-forwarding to a local custom port like

kubectl -n flyte port-forward service/minio <custom>:9000

Then, the

storage

section should let you override the Helm config for pyflyte

Yeah sorry I forgot to mention that ideally it should be possible without kubectl /: But I'm going to try another idea today. Long story short. this topic should be solved anyway properly by opening the servers ports on the long run...