Run-script for each component/service in cluster

[nklsla]

Create a separate shell script for each componentent. Each component should have a startup and shutdown script.

• [x] Bare minimum cluster componentes
• [x] Kubeflow
• [x] Grafana
• [x] Prometheus
• [x] Metrics
• [x] Docker image registry
• [x] NFS/Storage
• [x] Dynamic NFS provisioner (PVC's)
• [x] GPU-operator
• [x] Full startup, from initiated
• [x] Full shutdown, fix order
• [x] Shutdown to minimum cluster

[nklsla]

Need to wait for metric server before helm-installs

[nklsla]

Final shutdown seems to be to just SSH-shutdown the nodes