Add option for volume mounts

[jacobtomlinson]

It would be great to be able to configure volume mounts on the notebook and workers. This way if they need to access some network resource or other Kubernetes volume type they can.

Inspired by #75

[tonycpsu]

@jacobtomlinson : I see this issue is still open. Was this implemented properly in #102? Running the latest release (2021.4.0), I tried adding mounts: sections to my daskhub values.yaml file, but the mounts aren't showing up in the pods.

Relevant sections of my config:

jupyter:
  mounts:
    volumes:
      - name: project
        persistentVolumeClaim:
            claimName: claim-name
    volumeMounts:
        - name: data
          mountPath: /data


worker:
  mounts
    volumes:
      - name: project
        persistentVolumeClaim:
            claimName: claim-name
    volumeMounts:
        - name: data
          mountPath: /data

[jacobtomlinson]

@tonycpsu that looks right to me.

Have you tried running helm template --debug or helm get manifest and inspected the manifests which were rendered to see if the mounts are correct there?

[tonycpsu]

Not yet, no. I confess to being a bit of a Helm noob, so I've been using the Lens IDE to do my chart installs / upgrades instead of running helm from my shell.

I just ran helm -n [namespace] get manifest [releasename] but am not sure what I'm looking for in the output. I don't see any reference to anything from the mounts: section, but I also don't see any references to the dynamic home directory mount defined in singleuser: storage: that is working fine.

I'm not sure what the right invocation for helm template --debug is. I tried helm -n [namespace] template dask/daskhub --debug but am getting this:

Error: execution error at (daskhub/charts/jupyterhub/templates/proxy/deployment.yaml:28:32): Proxy token must be a 32 byte random string generated with `openssl rand -hex 32`!

[jacobtomlinson]

I think for the template debug you need to also pass your config file.

[tonycpsu]

@jacobtomlinson: thanks, found the -f [values.yaml] option for helm template, and not seeing the mounts in the output.

[consideRatio]

Using helm template ./dask -f test.yaml where test.yaml was like below, I conclude that passing values like this works as intended and that this was indeed closed by #102.

In the YAML below, I corrected a missing : after worker.mounts and I corrected incorrect k8s specification of declaring that the volumeMounts field name should point to the name of the declared volume.

jupyter:
  mounts:
    volumes:
      - name: project
        persistentVolumeClaim:
            claimName: claim-name
    volumeMounts:
      - name: project
        mountPath: /data


worker:
  mounts:
    volumes:
      - name: project
        persistentVolumeClaim:
            claimName: claim-name
    volumeMounts:
      - name: project
        mountPath: /data