Error creating hub and proxy pods

JupyterHub Zero to JupyterHub on Kubernetes
1

Hi All,

I’m following the steps in the zero to juypter hub for kubernetes guide (Installing JupyterHub — Zero to JupyterHub with Kubernetes documentation)
When we try to run step 3: kubectl get pod --namespace democlass03

we get
Updated property [core/project].
Fetching cluster endpoint and auth data.
kubeconfig entry generated for democlass03.
“jupyterhub” already exists with the same configuration, skipping
Hang tight while we grab the latest from your chart repositories…
…Successfully got an update from the “ngshare” chart repository
…Successfully got an update from the “jupyterhub” chart repository
Update Complete. ⎈Happy Helming!⎈
jhub-democlass03
democlass03
namespace/democlass03 created
Release “jhub-democlass03” does not exist. Installing it now.
Error: unable to build kubernetes objects from release manifest: [resource mapping not found for name: “user-placeholder” namespace: “” from “”: no matches for kind “PodDisruptionBudget” in version “policy/v1beta1”
ensure CRDs are installed first, resource mapping not found for name: “user-scheduler” namespace: “” from “”: no matches for kind “PodDisruptionBudget” in version “policy/v1beta1”
ensure CRDs are installed first]
No resources found in democlass03 namespace.
No resources found in democlass03 namespace.

and doesn’t create the hub and proxy pod.

Thanks for assistance,
David

2

Which version of Z2JH and K8s are you using? Can you share your Z2JH config with us?

3

hi Simon,

We figured out the “PodDisruptionBudget” in version “policy/v1beta1” error, it was due to the version of helm (0.11.1) and the GCP kubernetes version (1.25.4 on the server side and 1.26.0 on the client.) which no longer supported policy/v1beta1

We tried updating to helm version 1.0.0 however the issue is now JupyterHub doesn’t start, it attempts to start but then times out after 300 seconds. When i run the kubectl debug pod command during the Jupyterhub pod startup i get:

Defaulting debug container name to debugger-8xgbw.
Error from server (NotFound): pods “” not found

Also user-placeholder-0, 1 and 2 stays in pending status

user-placeholder-0 0/1 Pending 0 13h
user-placeholder-1 0/1 Pending 0 13h
user-placeholder-2 0/1 Pending 0 13h

when i run describe pod for each of the user-placeholders Events is set to , isn’t providing info on pending status

We’re running this on GCP.

Let me know if you need additional info.

Thanks,
David

4

Do you mean the version of the Z2JH Helm Chart or the version of Helm? Assuming it’s the former, that’s very old. Even 1.0.0 is quite old, the latest Z2JH release is 2.0.0. Can you try that, and if you still have problems please share your Z2JH config.

5

hi Simon,

upgrading to version 2 of z2h fixed it

Thanks,
David