tminaorg/jellyfin-kubernetes

Jellyfin setup with scalable transcoding in Kubernetes

UNMAINTAINED

This repo will be archived when I finish work on morana

jellyfin-kubernetes

This is a Helm chart repo for jellyfin-rffmpeg and jellyfin-rffmpeg-intro-skipper

Kubernetes

On Kubernetes you can use Longhorn RWX volumes (NFSv4) and mount said paths to Jellyfin host and workers (must be exactly the same mount points!).

You will need to generate an ssh key pair and add them as a secret using the commands below. Make sure this is done in the same namespace as the helm install.

ssh-keygen -N '' -f ./id_rsa
kubectl create secret generic rffmpeg-ssh --from-file=private_key=./id_rsa --from-file=public_key=./id_rsa.pub

jellyfin-kubernetes's Issues

Hardware Transcoding

Looking at https://github.com/aleksasiriski/jellyfin-kubernetes/blob/main/values.yaml#L70 this chart seems to do a more direct way of loading a gpu rather than using a device plugin as seen in the docs. Can we add support for a custom limits section for the deployments? I could also see benefits in using it for CPU / ram limits as well as assigning the GPU.