-
Notifications
You must be signed in to change notification settings - Fork 639
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
support for MicroShift/OpenShift 4.15+ through helm chart, static manifest and time-slicing manual tests #702
Conversation
Testing with Azure VM: MicroShift tested release:
Manifest file:
Verification POD:
Testing:
|
eb3eb94
to
718b6c2
Compare
@arthur-r-oliveira thanks for the contribution. The static deployments are around mostly for legacy purposes and not something we test regularly. The recommended mechanism for deploying the plugin is using the provided Helm charts. Would you be able to provide documentation on how to deploy the device plugin using helm for your target use case instead. If there is functionality missing that prevents this we can address any shortcomings. |
@elezar thanks for the heads-up! I'll give another with the helm charts and get back to you shortly. |
0a3cb48
to
48cf250
Compare
35fe28a
to
b74b0f1
Compare
aa00979
to
80384dc
Compare
@elezar I've closed this PR as have created much noise here, but will dispatch a second with small changes for deployments/helm/nvidia-device-plugin/templates/role-binding.yml and deployments/helm/nvidia-device-plugin/templates/role.yml. As you can see, original helm chart doesn't have appropriate Pod Security configurations in place to allow running nvidia-device-plugin with MicroShift 4.15+/k8s 1.28+
With the fixes:
To follow. |
Adding deployments/static/nvidia-device-plugin-privileged-with-service-account-and-time-slicing.yml, with time-slicing configuration inspired by GPU Operator https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/latest/gpu-sharing.html but for microshift support.