GKE resource mapping not found

[sachin-gopebble]

I am running testing out the GKE demo

• step 1. The script demo/clusters/gke/create-cluster.sh to create a GKE cluster
• step 2. The script demo/clusters/gke/install-dra-driver-gpu.sh to install dra driver

While applying the demo/specs/mig+mps/sharing-demo-parameters.yaml spec, I got the below errors. I suppose the CRDs are install by the install-dra-drive-gpu.sh script which includes the helm command

resource mapping not found for name: "gpu-ts-sharing" namespace: "sharing-demo" from "demo/specs/mig+mps/sharing-demo-parameters.yaml": no matches for kind "GpuClaimParameters" in version "gpu.resource.nvidia.com/v1alpha1"
ensure CRDs are installed first
resource mapping not found for name: "gpu-mps-sharing" namespace: "sharing-demo" from "demo/specs/mig+mps/sharing-demo-parameters.yaml": no matches for kind "GpuClaimParameters" in version "gpu.resource.nvidia.com/v1alpha1"
ensure CRDs are installed first
resource mapping not found for name: "mig-ts-sharing" namespace: "sharing-demo" from "demo/specs/mig+mps/sharing-demo-parameters.yaml": no matches for kind "MigDeviceClaimParameters" in version "gpu.resource.nvidia.com/v1alpha1"
ensure CRDs are installed first
resource mapping not found for name: "mig-mps-sharing" namespace: "sharing-demo" from "demo/specs/mig+mps/sharing-demo-parameters.yaml": no matches for kind "MigDeviceClaimParameters" in version "gpu.resource.nvidia.com/v1alpha1"
ensure CRDs are installed first

I also noticed that the deployments/helm/nvidia-dra-driver-gpu helm chart which get installed doesn't has any CRDs that create the resources related to gpu.resource.nvidia.com

[klueska]

It's very likely that this demo script is out of date. I'd be happy to look at a contribution to get it working again.

[jgehrcke]

Let's close this for now -- we can always come back to specific problems with new, focused tickets.