Installing Kubeflow using Bootstrapper
https://www.kubeflow.org/docs/started/getting-started-minikube/
Deploy Kubeflow to Kubernetes in Ubuntu
https://www.ubuntu.com/ai/install
Kubeflow uses Ksonnet to deploy the required components.
Jupyter Hub - for collaborative & interactive training A TensorFlow Training Controller with native distributed training A TensorFlow Serving for hosting Argo for workflows SeldonCore for complex inference and non TF models Ambassador for Reverse Proxy
kubectl get ns
kubectl get svc
kubectl -n kubeflow get svc -o wide
kubectl -n kubeflow get pod
kubectl logs -p -n kubeflow tf-hub-0
sudo rm -fr my-kubeflow
#----------
sudo snap disable microk8s
sudo snap enable microk8s
sudo microk8s.enable dns dashboard
sudo microk8s.enable storage
sudo iptables -P FORWARD ACCEPT
# forward ambassador service to from port 80 to 8080. we can access it from port 8080
POD=`kubectl -n kubeflow get pods --selector=service=ambassador | awk '{print $1}' | tail -1`
kubectl -n kubeflow port-forward $POD 8080:80 2>&1 >/dev/null &
# forward tf-hub service which contain juper notebook
POD=`kubectl -n kubeflow get pods --selector=app=tf-hub | awk '{print $1}' | tail -1`
kubectl -n kubeflow port-forward $POD 8000:8000 2>&1 >/dev/null &
http://localhost:8000/hub/login
# ACCESSING THE DASHBOARD
http://localhost:8080/api/v1/namespaces/kube-system/services/https:kubernetes-dashboard:/proxy/
http://localhost:8080/api/v1/namespaces/kube-system/services/http:monitoring-grafana:/proxy/
POD=`kubectl -n kube-system get pods --selector=k8s-app=influxGrafana | awk '{print $1}' | tail -1`
kubectl -n kube-system port-forward $POD 8011:8000 2>&1 >/dev/null &Set up a ksonnet application
ks init
Generate components for your app (e.g. Redis)
ks generate (alias of ks prototype use)
Deploy to a cluster
ks apply
Delete resources running on a cluster
ks deletea container-native workflow engine for Kubernetes