I am wondering if anybody knows the logical reasoning behind the syntax to remove a label in Kubernetes?
If I have a Pod named nginx labeled env=testing, I know that I can remove the label like so:
kubectl label pod nginx env-
But why do we need to use a -, not have some other command or a --remove flag to do so?
I think it is a design decision.
You can check the implementation here:
https://github.com/kubernetes/kubectl/blob/master/pkg/cmd/label/label.go#L418
func parseLabels(spec []string) (map[string]string, []string, error) {
labels := map[string]string{}
var remove []string
for _, labelSpec := range spec {
if strings.Contains(labelSpec, "=") {
parts := strings.Split(labelSpec, "=")
if len(parts) != 2 {
return nil, nil, fmt.Errorf("invalid label spec: %v", labelSpec)
}
if errs := validation.IsValidLabelValue(parts[1]); len(errs) != 0 {
return nil, nil, fmt.Errorf("invalid label value: %q: %s", labelSpec, strings.Join(errs, ";"))
}
labels[parts[0]] = parts[1]
} else if strings.HasSuffix(labelSpec, "-") {
remove = append(remove, labelSpec[:len(labelSpec)-1])
} else {
return nil, nil, fmt.Errorf("unknown label spec: %v", labelSpec)
}
}
for _, removeLabel := range remove {
if _, found := labels[removeLabel]; found {
return nil, nil, fmt.Errorf("can not both modify and remove a label in the same command")
}
}
return labels, remove, nil
}