rke2-capd: workload cluster fails, mgmt-cluster-state-values is disabled in the mgmt example
Summary
The rke2-capd management example disables the mgmt-cluster-state-values unit, but workload clusters need that configmap, so you can't deploy a workload cluster with the two rke2-capd examples. Seen on 1.7.7 and 1.8.0. Related to #4844 .
Steps to reproduce
- Deploy a management cluster from environment-values/rke2-capd (we had to disable rancher too, separate issue)
- Copy environment-values/workload-clusters/rke2-capd, set cluster_virtual_ip
- Run apply-workload-cluster.sh with it
What is the current bug behavior?
The workload sylva-units HelmRelease never gets ready and apply-workload-cluster.sh times out after 5 minutes.
What is the expected correct behavior?
Workload cluster gets deployed.
Relevant logs and/or screenshots
HelmRelease/sylva-units state changed: ValuesError - could not resolve ConfigMap chart values reference 'my-wc/mgmt-cluster-state-values' with key 'values': configmaps "mgmt-cluster-state-values" not found
Command timeout exceeded: HelmRelease/my-wc/sylva-units did not become Ready after 5m0sPossible fixes
environment-values/rke2-capd/values.yaml:11-12 disables the unit, while environment-values/workload-clusters/base/kustomization.yaml:25-27 needs the configmap. kubeadm-capd doesn't disable it.
Removing those two lines from environment-values/rke2-capd/values.yaml fixes it for us, we'll send an MR.