Nvidia deepops kubernetes GPU monitoring helm charts not working k8s version 1.17.6

Accelerated Computing DGX Systems (Data Center)
1

Dear team,

deepops deployed k8s cluster GPU monitoring helm chart is not working on k8s version 1.16.x and above. it is working only with k8s version 1.15.x.

Command used: (K8s version 1.17.6 deployed using DeepOPS)
helm install gpu-helm-charts/prometheus-operator --name prometheus-operator --namespace monitoring

Error mesage:
Error: validation failed: unable to recognize “”: no matches for kind “Deployment” in version “apps/v1beta1”

Following are the URL’s we are following to deploy GPU monitoring helm chart to deploy GPU monitoring stack with grafana and Prometheus.

Following are the URL’s we are using:

Section 3.2.3
and

Is it possible to deploy the monitoring stack helm charts in Kubernetes version 1.16.x and 1.17.x?

If yes kindly share the relevant documentation links …

Best Regards,
C Raja

2

Hey C Raja - are you still experiencing this issue? Deploying Kubernetes with DeepOps v20.08.1 and using Helm 3 to deploy the monitoring chart should work. If you still have the issue, please let me know what version of helm you’re using for the client and whether the tiller is deployed.