README.md

KubeCon EU 2018

This repository contains the demo code for my KubeCon EU 2018 talk about automating GPU infrastructure for Kubernetes on Container Linux.

Prerequisites

You will need a Google Cloud account with available quota for NVIDIA GPUs.

Getting Started

Edit the require.tf Terraform file and uncomment and add the details for your Google Cloud project:

$EDITOR require.tf

Modify the provided terraform.tfvars file to suit your project:

$EDITOR terraform.tfvars

Running

1. create cluster:

   terraform apply --auto-approve

2. get nodes:

   export KUBECONFIG="$(pwd)"/assets/auth/kubeconfig
   watch -n 1 kubectl get nodes

3. create GPU manifests:

   kubectl apply -f manifests

4. check status of driver installer:

   kubectl logs $(kubectl get pods -n kube-system | grep nvidia-driver-installer | awk '{print $1}') -c modulus -n kube-system -f

5. check status of device plugin:

   kubectl logs $(kubectl get pods -n kube-system | grep nvidia-gpu-device-plugin | awk '{print $1}' | head -n1 | tail -n1) -n kube-system -f

6. verify worker node has allocatable GPUs:

   kubectl describe node $(kubectl get nodes | grep worker | awk '{print $1}')

7. let's inspect the GPU workload:

   less manifests/darkapi.yaml

8. let's see if the GPU workload has been scheduled:

   watch -n 2 kubectl get pods
   kubectl logs $(kubectl get pods | grep darkapi | awk '{print $1}') -f

9. for fun, let's test the GPU workload:

   export INGRESS=$(terraform output | grep ingress_static_ip | awk '{print $3}')
   ~/code/darkapi/client http://$INGRESS/api/yolo

10. finally, let's clean up:

    terraform destroy --auto-approve

Projects Leveraged In This Demo

Component	URL
Kubernetes installer	https://github.com/poseidon/typhoon
GPU driver installer	https://github.com/squat/modulus
Kubernetes device plugin	https://github.com/kubernetes/kubernetes/blob/master/cluster/addons/device-plugins/nvidia-gpu/daemonset.yaml
sample workload	https://github.com/squat/darkapi