Skip to main content
NetApp Solutions

Kubeflow Deployment

Contributors kevin-hoke mboglesby
PDFs

This section describes the tasks that you must complete to deploy Kubeflow in your Kubernetes cluster.

Prerequisites

Before you perform the deployment exercise that is outlined in this section, we assume that you have already performed the following tasks:

  1. You already have a working Kubernetes cluster, and you are running a version of Kubernetes that is supported by the Kubeflow version that you intend to deploy. For a list of supported Kubernetes versions, refer to the dependencies for your Kubeflow version in the official Kubeflow documentation.

  2. You have already installed and configured NetApp Trident in your Kubernetes cluster. For more details on Trident, refer to the Trident documentation.

Set Default Kubernetes StorageClass

Before you deploy Kubeflow, we recommend designating a default StorageClass within your Kubernetes cluster. The Kubeflow deployment process may attempt to provision new persistent volumes using the default StorageClass. If no StorageClass is designated as the default StorageClass, then the deployment may fail. To designate a default StorageClass within your cluster, perform the following task from the deployment jump host. If you have already designated a default StorageClass within your cluster, then you can skip this step.

  1. Designate one of your existing StorageClasses as the default StorageClass. The example commands that follow show the designation of a StorageClass named ontap-ai-flexvols-retain as the default StorageClass.

Note The ontap-nas-flexgroup Trident Backend type has a minimum PVC size that is fairly large. By default, Kubeflow attempts to provision PVCs that are only a few GBs in size. Therefore, you should not designate a StorageClass that utilizes the ontap-nas-flexgroup Backend type as the default StorageClass for the purposes of Kubeflow deployment. 
$ kubectl get sc
NAME                                PROVISIONER             AGE
ontap-ai-flexgroups-retain          csi.trident.netapp.io   25h
ontap-ai-flexgroups-retain-iface1   csi.trident.netapp.io   25h
ontap-ai-flexgroups-retain-iface2   csi.trident.netapp.io   25h
ontap-ai-flexvols-retain            csi.trident.netapp.io   3s
$ kubectl patch storageclass ontap-ai-flexvols-retain -p '{"metadata": {"annotations":{"storageclass.kubernetes.io/is-default-class":"true"}}}'
storageclass.storage.k8s.io/ontap-ai-flexvols-retain patched
$ kubectl get sc
NAME                                 PROVISIONER             AGE
ontap-ai-flexgroups-retain           csi.trident.netapp.io   25h
ontap-ai-flexgroups-retain-iface1    csi.trident.netapp.io   25h
ontap-ai-flexgroups-retain-iface2    csi.trident.netapp.io   25h
ontap-ai-flexvols-retain (default)   csi.trident.netapp.io   54s

Kubeflow Deployment Options

There are many different options for deploying Kubeflow. Refer to the official Kubeflow documentation for a list of deployment options, and choose the option that is the best fit for your needs.

Note For validation purposes, we deployed Kubeflow 1.7 using deployKF 0.1.1.