Comments (5)
You should be able to experiment with https://github.com/intel/sriov-network-device-plugin where the resourceList
is be configured to expose your graphics card to a vfio device.
I've not tried with kubevirt but I'm using a similar mechanism to isolate (QAT) devices to Kata Containers backed pods.
from intel-device-plugins-for-kubernetes.
Thanks! I gave it a try, but I'm not entirely sure whether this is to work with virtualized GPUs, or would allow me to pass through an individual (physical) GPU?
I have created two vGPUs, which show up as /dev/vfio/9
and /dev/vfio/10
.
When I deploy the daemonset for the sriov-network-device-plugin, only one device is listed.
Creating a pod which uses this device fails: Error: failed to start container "appcntr1": Error response from daemon: linux runtime spec devices: error gathering device information while adding custom device "/dev/vfio/1": no such file or directory
.
Here's the configuration file:
{
"resourceList":
[
{
"resourceName": "gpu",
"rootDevices": [ "00:02.0" ],
"sriovMode": false,
"deviceType": "vfio"
}
]
}
And the output from describe pod:
I1205 13:36:09.720121 15 server.go:85] Allocate() called with &AllocateRequest{ContainerRequests:[&ContainerAllocateRequest{DevicesIDs:[0000:00:02.0],}],}
I1205 13:36:09.720239 15 vfioPool.go:76] vfio GetDeviceSpecs() called
I1205 13:36:09.720255 15 vfioPool.go:54] vfio GetEnvs() called
I1205 13:36:09.720272 15 vfioPool.go:70] vfio GetMounts() called
I1205 13:36:09.720279 15 server.go:94] AllocateResponse send: &AllocateResponse{ContainerResponses:[&ContainerAllocateResponse{Envs:map[string]string{PCIDEVICE_INTEL_COM_GPU: 0000:00:02.0,},Mounts:[],Devices:[&DeviceSpec{ContainerPath:/dev/vfio/vfio,HostPath:/dev/vfio/vfio,Permissions:mrw,} &DeviceSpec{ContainerPath:/dev/vfio/1,HostPath:/dev/vfio/1,Permissions:mrw,}],Annotations:map[string]string{},}],}
Let me know if missed something.
from intel-device-plugins-for-kubernetes.
@qmfrederik I was thinking about GVT-d so my idea won't work, sorry.
The link to your implementation points to GVT-g wiki.
from intel-device-plugins-for-kubernetes.
@mythi Makes sense, thanks.
The link should point to https://github.com/kubedroid/intel-device-plugins-for-kubernetes/blob/vgpu/cmd/vgpu_plugin/vgpu_plugin.go, I've updated the comment.
from intel-device-plugins-for-kubernetes.
Inactive for several years, if still valid please reopen.
from intel-device-plugins-for-kubernetes.
Related Issues (20)
- e2e-sgx: admissionwebhook tests are not run anymore
- [QAT] Why kernel driver passthrough all uio files HOT 9
- QAT testing/deployment updates HOT 3
- GPU label not added to nodes on Talos because i915 is built into the kernel HOT 1
- Publish accel-config-demo image to docker hub HOT 2
- intel-idxd-config-initcontainer cannot use UBI base images. HOT 5
- SGX-enabled pods sometimes get created without SGX device mounted HOT 5
- [QAT] No devices found in container without privilege HOT 2
- move openssl-qat-engine image to Ubuntu 24.04
- Intel 14th-gen Meteor Lake HOT 2
- Not compatible device: version HOT 2
- workflow: upgrade codecov upload method
- Prepare 0.30.0 release HOT 22
- Status of this Project HOT 3
- qat: deprecated and remove `-mode=kernel` and intel-qat-plugin-kerneldrv.Dockerfile
- Options for GPU Sharing between Containers Running on a Workstation HOT 5
- How to use the allocation policy HOT 2
- DSA and IAA demos fail with "mmap: Operation not permitted" HOT 6
- drop kube-rbac-proxy HOT 1
- Pod cannot detect Arc A380 GPU HOT 13
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from intel-device-plugins-for-kubernetes.