github/kata-containers
1
0
Fork 2
mirror of https://github.com/kata-containers/kata-containers.git synced 2026-04-09 05:23:43 +00:00
Code Issues Projects Releases Wiki Activity

Merge pull request #12799 from fitzthum/fixup-nv-doc-1

Browse Source 
docs: update flow for setting nvidia devices to ready
This commit is contained in:
Fabiano Fidêncio
2026-04-08 21:32:55 +02:00
committed by GitHub
parent 817580e35d 9119b4982c
commit 4561ae3e29
1 changed files with 5 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
docs/use-cases/NVIDIA-GPU-passthrough-and-Kata-QEMU.md
View File

@@ -213,12 +213,10 @@ API and kernel drivers, interacting with the pass-through GPU device.
An additional step is exercised in our CI samples: when using images from an
authenticated registry, the guest-pull mechanism triggers attestation using
Trustee's Key Broker Service (KBS) for secure release of the NGC API
authentication key used to access the NVCR container registry. As part of
this, the attestation agent exercises composite attestation and transitions
the GPU into `Ready` state (without this, the GPU has to explicitly be
transitioned into `Ready` state by passing the `nvrc.smi.srs=1` kernel
parameter via the shim config, causing NVRC to transition the GPU into the
`Ready` state).
authentication key used to access the NVCR container registry.
In this flow the CPU and all additional devices are attested.
GPUs will automatically be set to ready by NVRC per the NVRC configuration
flag in the default kernel command line.
## Deployment Guidance
@@ -465,8 +463,6 @@ kind: Pod
metadata:
name: cuda-vectoradd-kata
namespace: default
annotations:
io.katacontainers.config.hypervisor.kernel_params: "nvrc.smi.srs=1"
spec:
runtimeClassName: ${GPU_RUNTIME_CLASS_NAME}
restartPolicy: Never
@@ -613,12 +609,7 @@ You can author pod manifests leveraging your own containers, for instance,
containers built using the CUDA container toolkit. We recommend to start
with a CUDA base container.
The GPU is transitioned into the `Ready` state via attestation, for instance,
when pulling authenticated images. If your deployment scenario does not use
attestation, please refer back to the CUDA vectorAdd pod manifest. In this
manifest, we ensure that NVRC sets the GPU to `Ready` state by adding the
following annotation in the manifest:
`io.katacontainers.config.hypervisor.kernel_params: "nvrc.smi.srs=1"`
When using the GPU runtime classes, the GPUs will automatically be set to ready.
> **Notes:**
>