Unable to run rootless containers via Nvidia runtime

試著忘記壹切 提交于 2019-12-13 03:15:27

问题


I am totally new to the docker and nvidia. I am trying to install nvidia platform called clara on a unix server. While I try to run rootless containers via nvidia runtime, I get the below error (upon executing 2nd line of code). Can you please help me with it?

I already created a config file and placed this file under "config" folder and have updated this path in the 'Taskfile.yml' file. Moreover, through google search I found few threads which indicated that it can be due to cgroups. However, I created a Nvidia runtime hook file with the absolute path of the config.toml file. The content of the config.toml file is as shown below

When I execute the 2nd line of command, I get the error shown above

1)  ./run.sh default-docker-nokube  # this works fine and daemon 
                                               initialization is success
2)  docker -H unix://$XDG_RUNTIME_DIR/docker.sock run --runtime=nvidia -- 
    rm -it nvidia/cuda:10.0-devel nvidia-smi  #this command throws error

Please note that Nvidia hook file is placed under the bin folder of usernetes and am also attaching a screenshot of messages which are displayed under another window (server was running). As you can see, I received a message 'Daemon Initialization successful/complete' but after that when I try to run the 2nd line of code for client, I get the above error. I am just attaching the below screenshot only to help you debug the issue

I expect this command to be run successfully without any error. As I am very new, I am not sure how the output of this command would look like. So can you help me in executing this command successfully

来源:https://stackoverflow.com/questions/55676574/unable-to-run-rootless-containers-via-nvidia-runtime

易学教程内所有资源均来自网络或用户发布的内容,如有违反法律规定的内容欢迎反馈
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!