[Bug] Cannot create RKE2 cluster from yaml view

[jordojordo]

Setup

• Rancher version: v2.9-head (2b53406)
• Rancher UI Extensions: N/A
• Browser type & version: FF 125.0.3

Describe the bug

When attempting to provision an RKE2 cluster from the YAML view, the formatting of the generated yaml is incorrect, causing the creation to fail.

To Reproduce

• Install Rancher v2.9-head
• Create RKE2 cluster
• Select "Edit as YAML"
• Attempt to create the cluster

Result An error banner shows:

bad indentation of a mapping entry (13:56)

10 | ...
11 | ... t
12 | ...
13 | ... ustomization: overrideAffinity: {}
-----------------------------------------^
14 | ...
15 | ... ments: {}

Screenshots

Additional context

It appears that the schema returned is broken, there are a few instances with incorrect spacing and two duplicate entries.

• Incorrect spacing:
  • clusterAgentDeploymentCustomization: overrideAffinity: {}
  • fleetAgentDeploymentCustomization: overrideAffinity: {}
  • rkeConfig: chartValues:
• duplicate entries:
  • appendTolerations: []
  • overrideResourceRequirements: {}

Yaml returned for rke2 config

apiVersion: provisioning.cattle.io/v1
kind: Cluster
metadata:
  name: #string
  annotations:
    {}
    #  key: string
  labels:
    {}
    #  key: string
  namespace: fleet-default
spec:
  clusterAgentDeploymentCustomization: overrideAffinity: {}
  appendTolerations: []
  overrideResourceRequirements: {}
  defaultPodSecurityAdmissionConfigurationTemplateName: ''
  fleetAgentDeploymentCustomization: overrideAffinity: {}
  appendTolerations: []
  overrideResourceRequirements: {}
  kubernetesVersion: v1.28.9+rke2r1
  localClusterAuthEndpoint:
    caCerts: ''
    enabled: false
    fqdn: ''
  rkeConfig: chartValues:
    rke2-calico: {}
  upgradeStrategy:
    controlPlaneConcurrency: '1'
    controlPlaneDrainOptions:
      deleteEmptyDirData: true
      disableEviction: false
      enabled: false
      force: false
      gracePeriod: -1
      ignoreDaemonSets: true
      skipWaitForDeleteTimeoutSeconds: 0
      timeout: 120
    workerConcurrency: '1'
    workerDrainOptions:
      deleteEmptyDirData: true
      disableEviction: false
      enabled: false
      force: false
      gracePeriod: -1
      ignoreDaemonSets: true
      skipWaitForDeleteTimeoutSeconds: 0
      timeout: 120
  machineGlobalConfig:
    cni: calico
    disable-kube-proxy: false
    etcd-expose-metrics: false
    profile: null
  machineSelectorConfig:
    - config:
        protect-kernel-defaults: false
  etcd:
    disableSnapshots: false
    s3: null
    snapshotRetention: 5
    snapshotScheduleCron: 0 */5 * * *
  registries:
    configs: {}
    mirrors: {}
  machinePools: []
  machineSelectorConfig:
    - config: {}
#  agentEnvVars: string
#  cloudCredentialSecretName: string
#  clusterAPIConfig: string
#  defaultClusterRoleForProjectMembers: string
#  defaultPodSecurityPolicyTemplateName: string
#  enableNetworkPolicy: string
#  redeploySystemAgentGeneration: int
__clone: true

[richard-cox]

It looks like a lot of the spec props have lost their hierarchy

The schema looks ok, we should check value that goes into the yaml converter

[richard-cox]

This could be impacted by https://github.com/rancher/rancher/issues/45157, @MbolotSuse & @tomleb might be able to provide a debug image to validate that fix against this bug.

In the mean time we should still confirm the json object we convert to yaml is correct.

[gaktive]

Taking blocker off for now though this is important to fix before we release 2.9.0

[tomleb]

@richard-cox I'm working on https://github.com/rancher/rancher/issues/45157 now. I need to rebase my changes on top of @MbolotSuse 's fix, which was also affecting the schema definitions. Once that's done (by EOD) I'll ping you.

EDIT: To clarify, I don't know if the issue I'm working on causes what you see here.

[richard-cox]

Tom has kindly got an image ready for us to test with, check in with him for details