{{< listversions >}}

{{< listexamples >}}

Overview#

The set-namespace function update or add namespace to all namespaced
resources. Kubernetes supports multiple virtual clusters backed by the same
physical cluster through namespaces.

Namespaces are often used in the following scenarios:

• Separate resources between environments (prod, staging and test).
• Separate resources between different team or users to divide resource quota.

You can learn more about namespace [here][namespace].

Usage#

This function can be used with any KRM function orchestrators (e.g. kpt).

For all namespaced resurces, the set-namespace function adds the namespace
if metadata.namespace doesn't exist. Otherwise, it updates the existing value.
It will skip the resources that are known to be cluster-scoped (e.g. Node
, CustomResourceDefinitions, ClusterRole). Whether a resource is namespaced
is determined by the OpenAPI schema. If the API path
contains namespaces/{namespace} then the resource is considered namespaced.
Otherwise, it's not. Currently, this function is using API version 1.20.4.

In addition to updating the metadata.namespace field for applicable resources,
by default the function will also update the [fields][commonnamespace] that
target the namespace. There are a few cases that worth pointing out:

• If there is a Namespace resource, its metadata.name field will be updated.
• If there's a RoleBinding or ClusterRoleBinding resource, the function will
  update the namespace in the ServiceAccount if one of the following are true:
  1. the subject element name is default.
  2. the subject element name matches the name of a ServiceAccount resource declared in the package.

In the following example, the set-namespace function will update:

• subjects[0].namespace since subjects[0].name is default.
• subjects[1].namespace since subjects[1].name matches a ServiceAccount
  name declared in the package.

apiVersion: v1
kind: ServiceAccount
metadata:
  name: service-account
  namespace: original-namespace