VM GPU - DaoCloud Enterprise

[giscus[bot]]

VM GPU - DaoCloud Enterprise

DaoCloud Enterprise 5.0 is the next-generation PaaS platform, created with many leading cloud-native technologies, to easily promote and boost your digital business.

https://docs.daocloud.io/en/virtnest/vm/vm-gpu

[KyrieZhang1]

why my yaml don't create vm by kubevitr ?

vm.yaml:

apiVersion: kubevirt.io/v1
kind: VirtualMachine
metadata:
name: ubuntu-vm-ph-v2
spec:
runStrategy: Always
template:
metadata:
annotations:
kubevirt.io/selinux.context: "system_u:system_r:container_file_t:s0"
kubevirt.io/fsGroup: "107" # KubeVirt 默认组（通常为 qemu）
labels:
kubevirt.io/domain: ubuntu-vm-ph-v2
spec:
domain:
resources:
requests:
memory: 4Gi
cpu: 1
nvidia.com/GH100_H100_SXM5_80GB: 1 # 请求 2 个 GPU
limits:
memory: 4Gi
cpu: 1
nvidia.com/GH100_H100_SXM5_80GB: 1 # 最大 2 个 GPU
devices:
gpus:
- deviceName: nvidia.com/GH100_H100_SXM5_80GB
name: gpu1
- deviceName: nvidia.com/GH100_H100_SXM5_80GB
name: gpu2
disks:
- name: rootdisk
disk:
bus: virtio
- name: cloudinitdisk
disk:
bus: virtio
interfaces:
- name: default
masquerade: {}
machine:
type: q35
networks:
- name: default
pod: {}
volumes:
- name: rootdisk
persistentVolumeClaim:
claimName: dvv2 # 使用已有的 DataVolume 对应的 PVC
- name: cloudinitdisk
cloudInitNoCloud:
userData: |
#cloud-config
manage_etc_hosts: true
hostname: ubuntu-vm-ph-v2
ssh_pwauth: true
chpasswd:
expire: false
users:
- {name: root, password: 123456, type: text}
users:
- name: root
lock_passwd: false
sudo: ALL=(ALL) NOPASSWD:ALL
shell: /bin/bash

if [gpus.deviceName: nvidia.com/GH100_H100_SXM5_80GB only one] example：

    devices:
      gpus:
      - deviceName: nvidia.com/GH100_H100_SXM5_80GB
        name: gpu1

my vm normal

my node resoures:
[
"memory": "2112460648Ki",
"nvidia.com/GH100_H100_SXM5_80GB": "8"]