qingcloud-k8s-device-plugin's Introduction

Kubernets QingCloud Device Plugin

This is a kubernets device plugin which aims to help qingcloud user utilize GPU functionality with ease.

Following resources are implemented. if you need more resource please raise issue and let us know

name	ResourceName
GPU	nvidia.com/gpu

e.g.:

apiVersion: v1
kind: Pod
metadata:
  name: demo-pod
spec:
  containers:
    -
      name: demo-container-1
      image: dockerhub.qingcloud.com/google_containers/pause:2.0
      resources:
        limits:
          nvidia.com/gpu: 2 # requesting 2 GPU

Installation

The plugin is a daemon process and will listen to the unix socket under /var/lib/kubelet/device-plugins/

to launch the process you can simply run commands "qingcloud-gpu", other options are listed as below

./qingcloud-gpu -h
Device plugin to help kubernetes cluster utilize GPU resources on Qingcloud

Usage:
  qingcloud-gpu [flags]

Flags:
      --alsologtostderr                  log to standard error as well as files
      --container-path string            Path on the container that mounts '-host-path' (default "/usr/local/nvidia")
  -h, --help                             help for qingcloud-gpu
      --host-path string                 Path on the host that contains nvidia libraries. This will be mounted inside the container as '-container-path' (default "/home/kubernetes/bin/nvidia")
      --log_backtrace_at traceLocation   when logging hits line file:N, emit a stack trace (default :0)
      --log_dir string                   If non-empty, write log files in this directory
      --logtostderr                      log to standard error instead of files
      --stderrthreshold severity         logs at or above this threshold go to stderr (default 2)
  -v, --v Level                          log level for V logs
      --vmodule moduleSpec               comma-separated list of pattern=N settings for file-filtered logging