-
Story
-
Resolution: Unresolved
-
Major
-
None
-
None
-
None
-
Quality / Stability / Reliability
-
False
-
-
False
-
5
-
None
-
OCP Node Devices Sprint 283
DRA APIs have been enabled in OCP 4.21[1]
There has been some work done earlier to test the integration of DRAAPIs on an NVIDIA GPU in OCP-4.20 by Abu[2]
We have the NVIDIA Gpu operator available only till OCP-4.20 and not yet for OCP-4.21[3] in the software catalogue
So, the idea here is to deploy the nvidia dra driver via helm and test the downstream enabled DRA APIs on the NVIDIA GPU.
Reference to install the NVIDIA GPU Operator via helm:
DRA driver installation on Openshift: https://github.com/NVIDIA/k8s-dra-driver-gpu/pull/82/changes
SCC fix for Openshift: https://github.com/NVIDIA/k8s-dra-driver-gpu/pull/569
Slack conversation on #forum-dra [4]
[1] - https://github.com/openshift/api/pull/2498
[2] - https://github.com/openshift/origin/pull/29842
[3] - https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/latest/platform-support.html#supported-operating-systems-and-kubernetes-platforms
[4] - https://redhat-internal.slack.com/archives/C066VKUM5HP/p1768241259634119
- causes
-
OCPNODE-4044 Manual validation of downstream DRA APIs with an NVIDIA dra driver 25.12
-
- To Do
-
- links to