cannot load lib/helpers.star: Evaluating starlark template:

[anibal-aguila]

Hi, we have the following issue trying to setup a new cluster.

What steps did you take: tanzu cluster create tkgs-shared-base --file /root/downloads/tanzu-bundle-1.4/tkg-shared-base.yaml --tkr=v1.20.9---vmware.1-tkg.1.a4cee5b -v 6

What happened: Validating configuration... Error: failed to create Tanzu Kubernetes Cluster service for vSphere workload cluster:

• cannot load lib/helpers.star: Evaluating starlark template: in overlay.yaml:3 | #@ load("lib/helpers.star", "get_bom_data_for_tkr_name", "get_default_tkg_bom_data", "kubeadm_image_repo", "get_image_repo_for_component", "get_vsphere_thumbprint")

  reason:

  • NoneType has no .split field or method in get_tkr_version_from_tkr_name 15 | strs = tkr_name.split("---") in get_bom_data_for_tkr_name 42 | tkr_version = get_tkr_version_from_tkr_name(data.values.KUBERNETES_RELEASE) in

What did you expect: Cluster up and running

Anything else you would like to add: [Additional information that will assist in solving the issue.]

Environment:

• ytt version (use ytt --version): 0.34.0
• OS (e.g. from /etc/os-release): Photon 4.0

thanks in advance,

[pivotaljohn]

Hey @anibal-aguila.

So, the symptom appears to be that tkr_name is None (the Python/Starlark equivalent to nil or null).

Just based of the name of the function, it looks like tkr_name is getting its value from data.values.KUBERNETES_RELEASE.

Are you able to verify that that data value is being set?

• One clue is that the data value is written in ALL CAPS. I often see that when the data value is being set via an environment variable.
• that said, I also notice that the command line arg --tkr=v1.20.9---vmware.1-tkg.1.a4cee5b contains that triple-dash... perhaps the Tanzu CLI expects to place the value set for tkr to this KUBERNETES_RELEASE data value?

I see that you're using Tanzu. I suspect that you'd get additional help in the Kubernetes slack: https://kubernetes.slack.com/archives/C02GY94A8KT. Visit http://slack.k8s.io/ to join the Kubernetes Slack workspace if you are not already there.

[pivotaljohn]

We'll close this issue for now. @anibal-aguila, if you would like more assistance, please feel free to comment here; we'll see it and re-open this issue. Hopefully we haven't heard from you because the issue is resolved. 🤞🏻