Replies: 3 comments 5 replies
-
|
Probably found out: |
Beta Was this translation helpful? Give feedback.
-
|
Ok, the previous patch seems to be accepted, but cannot say if it really went through: I still cannot run workloads on those nodes without setting the runtimeClassName in the worload spec. So, is the patch ok? How can I be sure it went through, letting me run workloads without having to always set the runtimeClassName? Thank you. |
Beta Was this translation helpful? Give feedback.
-
|
Hi, So I think I was confused by the docs, stating the patch would make the runtimeClass the default one: https://docs.siderolabs.com/talos/v1.7/configure-your-talos-cluster/hardware-and-drivers/nvidia-gpu#optional-setting-the-default-runtime-class-as-nvidia Thank you for the insight, there's probably some clearing on the docs to do? |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Hi,
Trying to run gpu workloads on Omni managed cluster.
The needed patches work, and I can safely run pods defining the runtimeClass in the workload yaml.
Thing is: I cannot seem to be able to patch the default runtime on the node:
Is there a different way to make it work?
Thank you.
Beta Was this translation helpful? Give feedback.
All reactions