Hello all. I'm trying to use Rancher to install Helm Chart Repos on a connected cluster where the master nodes on that cluster must use a proxy for internet based request. The Helm chart repo times out trying to download the

index.yaml

. I've added my proxy config to

/etc/default/k3s

but that does not seem to affect the Helm Chart install from Rancher. Any idea what I am missing here or where to configure my proxy for this use case? FWIW, I can install the same repo locally on my Rancher cluster but those nodes do not require the use of a proxy for internet request.

You can get around this using the HelmChart CRD and apply the chart in base64 https://docs.rke2.io/helm/

That would seem to make chart updates a manual process, No offense but not a fan of that answer. I'd like to think that if Rancher is going to install the Helm Chart repo on the remote cluster, then the remote cluster should honor the environment variables for proxy either from the OS or from the K3s env file

/etc/default/k3s

. Perhaps if I had a technical breakdown of what Rancher is doing to execute these actions, I could figure out what is missing and where to make it honor using the proxy. I added these proxy env var's recently so maybe it's just a pod that needs to be restarted to pick those up but I don't know which pods (if any) Rancher is talking to on the remote cluster to go get this repo's index.yaml file.

Rancher's helm chart has no-proxy configuration options as well

Do you know what is executing the equivalent

helm repo add ...

command on the remote cluster side? Something has to be making that call and failing to pull it because it's not routing through my proxy.

gitjob i think, part of fleet

I'll have to look into that. Not 100% clear what that entails but I will do the research and check on that suggestion. Thank you for your guidance.