deep-learning-toolkit

Kubernetes

For connecting DLTK to a Kubernetes environment, please use the kubernetes Connector in the Create Environment dialog. Depending on the type of Kubernetes environment you must also specify additional fields. This guide describes how this works for typical Kubernetes distributions.

If you want to create a new Environment, please see the Connect to Environments documentation.

Docker Desktop

Docker Desktop is an application for MacOS and Windows machines. It ships with a fully functional Kubernetes environment on your desktop.

If you want to learn more about Docker Desktop, please see the official Docker Desktop product documentation.

The following steps describe how to connect DLTK to Docker Desktop, running on the same host:

  1. Download and install Docker Desktop for Windows or Linux from https://www.docker.com/get-started
  2. Open the Docker Desktop settings window, select the Kubernetes page and check Enable Kubernetes
  3. Click the Apply & Restart button and wait until Kubernetes is running
  4. Open the kubeconfig file which is typically located in ~/.kube/config
  5. In the Create Environment dialog (see Connect to Environments) make sure the following fields are specified:
    • Set Auth Mode to cert-key
    • Set Cluster Url to https://kubernetes.docker.internal:6443
    • Set Client Cert to value of client-certificate-data in the kubeconfig
    • Set Cluster Ca to value of certificate-authority-data in the kubeconfig
    • Set Client Key to value of client-key-data in the kubeconfig
    • Set Ingress Mode to node-port
    • Set Node Port Url to http://localhost
    • Set Storage Class to hostpath

MicroK8s

MicroK8s is the smallest, fastest, fully-conformant Kubernetes that tracks upstream releases and makes clustering trivial. MicroK8s is great for offline development, prototyping, and testing. If you want to learn more about MicroK8s, please see the official MicroK8s product documentation.

The following steps describe how to connect DLTK to MicroK8s, running on the same host:

  1. Download and install MicroK8s for Windows, Linux or MacOS from the official MicroK8s site
  2. Perform the setup on the CLI to start MicroK8s
  3. Verify that MicroK8s is running and enable add-ons:
    • Enable dns and storage with microk8s enable dns storage
    • Optionally, if you have GPU accessible and properly configured you can enable gpu with microk8s enable gpu
    • Enable ingress with microk8s enable ingress and modify the ingress configmap using microk8s kubectl edit configmap -n ingress nginx-load-balancer-microk8s-conf with the following changes to be added to:
      • data:
        • ` proxy-body-size: “0”`
        • ` proxy-read-timeout: “100000”`
        • ` proxy-connect-timeout: “100000”`
        • ` ssl-redirect: “false”`
  4. Display the microk8sconfig with microk8s config and use to fill the information in the following step 5
  5. In the Create Environment dialog (see Connect to Environments) make sure the following fields are specified:
    • Set Auth Mode to user-password
    • Set Cluster Url to the value server of your cluster in the microk8sconfig, e.g. https://your-microk8s-host-or-ip-or-localhost:16443
    • Set Cluster Ca to value of certificate-authority-data in the microk8sconfig
    • Set Ingress Mode to ingress
    • Set Ingress Class to nginx
    • Set Ingress Url to the value server of your cluster in the microk8sconfig, e.g. http://your-microk8s-host-or-ip-or-localhost or https if you configure microk8s ingress with certificates
    • Set User Name to value of username of your user in microk8sconfig
    • Set User Password to value of password of your user in microk8sconfig

Amazon Elastic Kubernetes Service (EKS)

EKS is a fully managed Kubernetes service.

If you want to learn more about EKS, please see the official Amazon Elastic Kubernetes Service product documentation.

The following steps describe how to connect DLTK to EKS cluster:

  1. Set up and log into your AWS account
  2. Create a EKS cluster by following the Getting started with Amazon EKS guide
  3. Setup an Ingress Controller in your EKS cluster
  4. In the Create Environment dialog (see Connect to Environments) make sure the following fields are specified:
    • Set Auth Mode to aws-iam
    • Set Aws Access Key Id to your IAM Access Key
    • Set Aws Secret Access Key to IAM Access Secret
    • Set Aws Cluster Name to the name of your EKS cluster
    • Set Aws Region Name to the name of the AWS region the EKS cluster runs in
    • Set Ingress Mode to ingress
    • Set Ingress Url to the url of yor Ingress Controller
    • Set Storage Class to gp2

Azure Kubernetes Service (AKS)

If you want to learn more about AKS, please see the official Azure Kubernetes Service product documentation.

Google Kubernetes Engine (GKE)

If you want to learn more about GKE, please see the official Google Kubernetes Engine product documentation.

This site is open source. Improve this page.