-
Notifications
You must be signed in to change notification settings - Fork 470
Open
Labels
lifecycle/frozenIndicates that an issue or PR should not be auto-closed due to staleness.Indicates that an issue or PR should not be auto-closed due to staleness.
Description
We have two packages/groups defined in our APIs directory. There is a struct with the same name in each; one of which is intended to generate a CRD, and one of which is just a field within another CRD.
Example:
// in apis/ourapp/v1beta2/ourresource1_types.go
// +kubebuilder:object:root=true
// +kubebuilder:subresource:status
// +kubebuilder:storageversion
// +kubebuilder:resource:path=missioncontrolclusters,shortName=mccluster;mcclusters
type Bar struct {
metav1.TypeMeta `json:",inline"`
metav1.ObjectMeta `json:"metadata,omitempty"`
Spec BarSpec `json:"spec,omitempty"`
Status BarStatus `json:"status,omitempty"`
}
type BarSpec {
...
Foo *Foo `json:"foo,omitempty"`
...
}
type Foo struct {
...
}
// in apis/ourapp/v1alpha1/ourresource2_types.go
// +kubebuilder:object:root=true
// +kubebuilder:subresource:status
// +kubebuilder:storageversion
// +versionName=v1alpha1
// +kubebuilder:printcolumn:name="Error",type=string,JSONPath=".status.error",description="Latest reconcile error"
type Foo struct {
metav1.TypeMeta `json:",inline"`
metav1.ObjectMeta `json:"metadata,omitempty"`
Spec FooSpec `json:"spec,omitempty"`
Status FooStatus `json:"status,omitempty"`
}
When the CRDs get generated, we're seeing both a v1alpha1 and v1beta2 resource version being created within the CRD.
This is incorrect and results in the spec of the Foo being deleted due to some sort of conversion webhook mechanics.
Is there a documented constraint that a struct with the same name as a kubebuilder object root cannot exist in another API version/package?
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
lifecycle/frozenIndicates that an issue or PR should not be auto-closed due to staleness.Indicates that an issue or PR should not be auto-closed due to staleness.