【发布时间】:2020-09-21 10:45:15
【问题描述】:
我需要在 kubernetes 上运行一个 cuda 二进制文件。我已将节点设置为将kubernetes nvidia device plugin 与 nvidia-docker2 一起使用。这是我的 Dockerfile:
FROM ubuntu:18.04
COPY addarrays /addarrays
ENTRYPOINT [ "/addarrays" ]
当我通过 nvidia-docker2 或 kubernetes 运行 docker 映像时,会出现以下错误:
Cuda failure addarrays.cu:9: 'CUDA driver version is insufficient for CUDA runtime version'
Cuda failure addarrays.cu:9: 'CUDA driver version is insufficient for CUDA runtime version'
Cuda failure addarrays.cu:9: 'CUDA driver version is insufficient for CUDA runtime version'
Cuda failure addarrays.cu:9: 'CUDA driver version is insufficient for CUDA runtime version'
Cuda failure addarrays.cu:9: 'CUDA driver version is insufficient for CUDA runtime version'
Cuda failure addarrays.cu:9: 'CUDA driver version is insufficient for CUDA runtime version'
addarrays: addarrays.cu:62: int main(): Assertion `hostArrayTmp[i] == hostArrayDest[i]' failed.
看起来我的 docker 映像需要一个 nvidia 驱动程序。我已经像这样修改了 Dockerfile:
FROM ubuntu:18.04
RUN apt install software-properties-common
RUN add-apt-repository ppa:graphics-drivers
RUN apt update
RUN apt install nvidia-driver-440
COPY addarrays /addarrays
ENTRYPOINT [ "/addarrays" ]
安装 add-apt-repository 需要 software-properties-common,但安装失败并显示以下消息:
E: Unable to locate package software-properties-common
我需要做什么才能在我的 docker 映像中安装 nvidia 驱动程序?
【问题讨论】:
标签: docker kubernetes gpu nvidia nvidia-docker