Adding Contributors with the Users Secret

Add contributors to an AI/ML tenant-based Kubeflow profile.

Prerequisites

  • Required access rights: Kuberentes Administrator or Tenant Administrator

About this task

NOTE Kubeflow automatically generates a secret in the tenant namespace for users who are signed in to any KD Jupyter notebook in the ML Ops Tenant. For users who are not signed in, follow this procedure to generate a secret manually.

Contributors added to the AI/ML tenant-based Kubeflow profile gain access to the tenant namespace. In the Kubeflow Dashboard, the tenant namespace becomes available in the dropdown list of available namespaces.

Procedure

Create a users secret in the tenant namespace: 
apiVersion: v1
kind: Secret
type: Opaque
metadata:
  name: kubeflow-users-secret
  namespace: <tenant_namespace>
data:
  users: <base64_encoded_usernames>
For encoding purposes, use a comma to separate usernames. For example:
users: ZGV2MSxxYTEK # - dev1,qa1

After you create the users secret, Kubeflow creates a role binding for each user.

You can check the rolebindings to validate that a contributor was added successfully. For example:

kubectl get rolebindings -n <tenant_namespace> | grep <user_name>

 user-<user_name>-clusterrole-edit-cf4b9chf5t ClusterRole/kubeflow-edit 17s