Jupyter Hub on Dask using Internal LoadBalancer

12/12/2018

I wanted to ask you if anyone has experience with configuring Jupyterhub on DASK with Internal Load balancer? Default setup comes with External Preferably if Jupyter Hub's Load balancer would be mapped to a specific Internal IP or subnet? I'm trying to setup and environment which would not be accessible from outside.

I have managed to setup JupytherHub with Internal LoadBalancer previously by editing the config.yaml using this this repo for installation: https://jupyterhub.github.io/helm-chart/

proxy: 
  service:
    type: LoadBalancer
      annotations: {service.beta.kubernetes.io/aws-load-balancer-internal: "internal subnet"}

but this is not possible with dask.

Is there a way to achieve this or some another way? Any suggestions welcome!

Thank you

-- tregin
dask
dask-kubernetes

1 Answer

1/7/2019

Yes, this is possible.

I was able to do so with the following kubectl command:

kubectl annotate svc dask-scheduler cloud.google.com/load-balancer-type=Internal

After about 30 seconds the IP address will transfer to an internal one, if you have provisioned enough IP addresses for the sevices to utilise (you many need to allow more).

-- Chrisjw42
Source: StackOverflow