部署了一个模型服务
root@daiyu:/home/zhiman/vllm# kubectl get pod -n llm vllm-service-zhiman-vl-chat-75fd7f948f-59r6n -o yaml
apiVersion: v1
kind: Pod
metadata:
annotations:
hami.io/bind-phase: success
hami.io/bind-time: "1780883639"
hami.io/vgpu-devices-allocated: ;GPU-9b9030a3-d81f-7d9f-5d2b-fd98cdc57b2d,NVIDIA,6144,30:;
hami.io/vgpu-devices-to-allocate: ;;;
hami.io/vgpu-node: honglou
hami.io/vgpu-time: "1780883639"
如上hami.io/vgpu-devices-allocated: ;GPU 这个GPU前面多了一个;，下面的hami.io/vgpu-devices-to-allocate: ;;;也是出现了3个;
正常的如下：
root@daiyu:/home/zhiman# kubectl get pod -n llm mineru-new-7f8f6dc857-vfkkf -o yaml
apiVersion: v1
kind: Pod
metadata:
annotations:
hami.io/bind-phase: success
hami.io/bind-time: "1780723598"
hami.io/vgpu-devices-allocated: GPU-24aa6632-8f88-43ac-d286-0b7a65e1cee1,NVIDIA,24576,100:;
hami.io/vgpu-devices-to-allocate: ;
hami 2.9.0， hami-webui 1.2.0
现在导致webui上面不能显示到所有的GPU资源。
然后异常的配置
resources:
limits:
nvidia.com/gpu: "1"
nvidia.com/gpucores: "30"
nvidia.com/gpumem: "6144"
requests:
nvidia.com/gpu: "1"
nvidia.com/gpucores: "30"
nvidia.com/gpumem: "6144"
正常的配置
resources:
limits:
cpu: "16"
memory: 64Gi
nvidia.com/gpu: "1"
nvidia.com/gpucores: "100"
requests:
cpu: "1"
memory: 2Gi
nvidia.com/gpu: "1"
nvidia.com/gpucores: "100"
部署了一个模型服务
如上hami.io/vgpu-devices-allocated: ;GPU 这个GPU前面多了一个;，下面的hami.io/vgpu-devices-to-allocate: ;;;也是出现了3个;
正常的如下：
hami 2.9.0， hami-webui 1.2.0
现在导致webui上面不能显示到所有的GPU资源。
然后异常的配置
resources:
limits:
nvidia.com/gpu: "1"
nvidia.com/gpucores: "30"
nvidia.com/gpumem: "6144"
requests:
nvidia.com/gpu: "1"
nvidia.com/gpucores: "30"
nvidia.com/gpumem: "6144"
正常的配置
resources:
limits:
cpu: "16"
memory: 64Gi
nvidia.com/gpu: "1"
nvidia.com/gpucores: "100"
requests:
cpu: "1"
memory: 2Gi
nvidia.com/gpu: "1"
nvidia.com/gpucores: "100"