Cluster.spec.rkeConfig.machinePools[]
RKEMachinePool is the configuration for a RKE2/K3s machine pool within a provisioning cluster.
| Field / Type | Description |
|---|---|
autoscalingMaxSize | AutoscalingMaxSize is the autoscaler max node size, this maps to the autoscaling annotations which are mapped to the CAPI machineDeployment object
|
autoscalingMinSize | AutoscalingMinSize is the autoscaler min node size, this maps to the autoscaling annotations which are mapped to the CAPI machineDeployment object
|
cloudCredentialSecretName | CloudCredentialSecretName is the id of the secret used to provision the cluster. This field must be in the format of "namespace:name". NOTE: this field overrides the field of the same name on the cluster spec, allowing individual machine pools to use separate credentials.
|
controlPlaneRole | ControlPlaneRole defines whether the machines provisioned by this pool should be controlplane nodes. |
displayName | DisplayName is the display name for the generated CAPI machinedeployment object. Deprecated: this field is currently unused and will be removed in a future version.
|
drainBeforeDelete | DrainBeforeDelete defines whether the machines provisioned by this pool should be drained prior to deletion. |
drainBeforeDeleteTimeout | DrainBeforeDeleteTimeout defines the timeout for draining the machines provisioned by this pool before deletion.
|
dynamicSchemaSpec | DynamicSchemaSpec is a copy of the dynamic schema object's spec field at the time the machine pool was created. Since rancher-machine based MachineTemplates are not api-versioned, this field is used to drop new fields added to the driver if it has been upgraded since initial provisioning. This allows node drivers to be upgraded without triggering a reconciliation of the provisioning cluster. NOTE: This field can only be removed if the "provisioning.cattle.io/allow-dynamic-schema-drop" annotation is present on the provisioning cluster object, otherwise it will be reinserted.
|
etcdRole | EtcdRole defines whether the machines provisioned by this pool should be etcd nodes. |
hostnameLengthLimit | HostnameLengthLimit defines the maximum length of the hostname for machines in this pool. For Windows nodes utilizing NETBIOS authentication, a maximum of 15 should be set to ensure all nodes adhere to the protocol's naming requirements.
|
Labels is a list of labels to apply to the machines created by the CAPI machine deployment.
| |
NodeConfig is a reference to a MachineConfig object that will be used to configure the machines provisioned by this pool. The NodeConfig object will, in turn, be used to create a corresponding MachineTemplate object for the generated machine deployment.
| |
MachineDeploymentAnnotations are the annotations to add to the generated machine deployment.
| |
MachineDeploymentLabels are the labels to add to the generated machine deployment.
| |
machineOS | MachineOS is the operating system of the machines provisioned by this pool. This is only used to designate linux versus windows nodes.
|
maxUnhealthy | MaxUnhealthy specifies the minimum number of unhealthy machines that a MachineHealthCheck can tolerate before remediating unhealthy machines.
|
name | Name is the internal name of the machine pool. The generated CAPI machine deployment will be a concatenation of the cluster name and the machine pool name which, if over 63 characters is truncated to 54 with a sha256sum appended.
|
nodeStartupTimeout | NodeStartupTimeout allows setting the maximum time for MachineHealthCheck to consider a Machine unhealthy if a corresponding Node isn't associated through a `Spec.ProviderID` field. The duration set in this field is compared to the greatest of: - Cluster's infrastructure ready condition timestamp (if available) - Control Plane's initialized condition timestamp (if available) - Machine's infrastructure ready condition timestamp (if available) - Machine's metadata creation timestamp Defaults to 10 minutes. If you wish to disable this feature, set the value explicitly to 0.
|
paused | Paused indicates that the machine pool is paused, preventing CAPI controllers from reconciling it. NOTE: this only applies to the corresponding generated machine deployment object, not the generated machines themselves. |
quantity | Quantity is the desired number of machines in the machine pool.
|
RollingUpdate is the configuration for the rolling update of the generated machine deployment.
| |
Taints is a list of taints to apply to the machines created by the CAPI machine deployment.
| |
unhealthyNodeTimeout | UnhealthyNodeTimeout specifies the maximum duration a generated MachineHealthCheck should wait before marking a not ready machine as unhealthy.
|
unhealthyRange | UnhealthyRange specifies the number of unhealthy machines in which a MachineHealthCheck is allowed to remediate.
|
Additional userdata to merge with rancher-defined userdata. Currently, this field is only supported for external CAPI infrastructure providers, and will have no effect for Rancher-provisioned nodes. Userdata for Rancher-provisioned nodes can be set in the machine configuration. WARNING: Do not include sensitive data, this field is visible in non-Secret API objects. EXPERIMENTAL: this field is experimental and future compatibility is not guaranteed.
| |
workerRole | WorkerRole defines whether the machines provisioned by this pool should be worker nodes. |