HAMi成功安装，服务运行正常，但是算力无法切分

[ak47947]

What happened: 使用GPU Operator安装Kubernetes GPU 环境搭建，然后安装HAMi插件，服务安装正常，但是GPU数量还是显示1，在容器中也未切分

What you expected to happen: GPU数量显示10份（默认），容器中资源得到限制

How to reproduce it (as minimally and precisely as possible): 使用GPU Operator安装Kubernetes GPU 环境

Anything else we need to know?:

1. 安装后服务正常

   

2. GPU没有切分

   

• The output of nvidia-smi -a on your host

  

• Your docker or containerd configuration file (e.g: /etc/docker/daemon.json) 配置无问题

• The hami-device-plugin container logs

  

• The hami-scheduler container logs

  

• The kubelet logs on the node (e.g: sudo journalctl -r -u kubelet)

• Any relevant kernel output lines from dmesg

Environment:

• HAMi version: v2.4.0
• nvidia driver or other AI device driver version: 见上图
• Docker version from docker version : Docker version 20.10.24, build 297e128
• Docker command, image and tag used
• Kernel version from uname -a : ubuntu 5.15.0-124-generic
• Others:

[archlitchi]

have you uninstalled nvidia-k8s-device-plugin before installing HAMi?

[ak47947]

have you uninstalled nvidia-k8s-device-plugin before installing HAMi?

已经安装了的，是否需要卸载

[ak47947]

我通过helm uninstall hami -n kube-system 卸载后重装hami解决了，现在可以看到GPU信息了

进入容器也可以看到隔离信息了

[ak47947]

发现一个新的问题，这个问题可能是因为开关机为主机增加和删除新的显卡引起的，在增加和删除显卡后，hami会失效