-
Notifications
You must be signed in to change notification settings - Fork 332
K8s, now what?
See the "Kubernetes on NVIDIA DGX Servers User Guide" for Kubernetes usage examples:
https://docs.nvidia.com/datacenter/kubernetes/kubernetes-on-dgx-user-guide/index.html
With JupyterHub you can create a multi-user Hub which spawns, manages, and proxies multiple instances of the single-user Jupyter notebook server.
Project Jupyter created JupyterHub to support many users. The Hub can offer notebook servers to a class of students, a corporate data science workgroup, a scientific research project, or a high performance computing group.
Label nodes where you want the notebook pods to run:
kubectl label nodes dgx01 gpu=true
Modify config/jupyterhub-config.yml
if needed and deploy with Helm:
kubectl create ns jh
helm repo add jupyterhub https://jupyterhub.github.io/helm-chart/
helm repo update
helm install jupyterhub/jupyterhub --name=jupyterhub --namespace=jh -f config/jupyterhub-config.yaml --set prePuller.enabled=false
Remove:
helm del --purge jupyterhub
You may have to install/delete/re-install if there are RBAC errors. You can also try deleting the proxy pod so it's re-created
If you make changes to the config file:
helm upgrade jupyterhub jupyterhub/jupyterhub --version=v0.6 -f config/jupyterhub-config.yaml