charts/charts/stable/intel-gpu-plugin/README_CONFIG.md.gotmpl
2021-03-18 13:11:24 +01:00

25 lines
840 B
Go Template

{{- define "custom.custom.configuration.header" -}}
## Custom configuration
{{- end -}}
{{- define "custom.custom.configuration" -}}
{{ template "custom.custom.configuration.header" . }}
### Node Feature Discovery
If your cluster runs [Node Feature Discovery](https://github.com/k8s-at-home/charts/blob/master/charts/node-feature-discovery), you can deploy the device plugin only on nodes with Intel GPU by specifying the desired `nodeSelector` or `affinity` in your values. For example (make sure to update to your exact feature label):
```yaml
affinity:
nodeAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
nodeSelectorTerms:
- matchExpressions:
- key: feature.node.kubernetes.io/pci-0300_8086.present
operator: In
values:
- "true"
```
{{- end -}}