deepin 15.8 安装nvidia-docker2
Tofloor
poster avatar
ljd3000
deepin
2018-12-03 18:30
Author
本帖最后由 ljd3000 于 2018-12-3 10:35 编辑

*需安装nvidia独立驱动

一、安装包仓库
curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
curl -s -L https://nvidia.github.io/nvidia-docker/debian8/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
sudo apt-get update

二、安装nvidia-dochker2, 重启docker服务
sudo apt-get install -y nvidia-docker2
sudo pkill -SIGHUP dockerd

*安装会覆盖更改/etc/docker/daemo.json,


在deepin 上还需要安装几个包

sudo apt-get install nvidia-opencl-common libnvidia-encode1 libnvidia-fbc1 libnvidia-ifr1 libgles1-nvidia libcuda1 nvidia-smi

*缺省会安装nvidia-smi i386版本,但在docker环境下会报"no such file or directory", 需安装amd64版本


修改/etc/nvidia-container-runtime/config.toml, 将ldconfig="@/sbin/ldconfig.real"注释掉

增加/etc/ld.so.config.d/nvidia.conf

/usr/lib/x86_64-linux-gnu/nvidia
/usr/lib/x86_64-linux-gnu/nvidia/current

执行ldconfig

*下面的命令可以显示调试信息,可以看出缺少的库,i386的库没关系

sudo nvidia-container-cli -k -d /dev/tty info

三、测试
docker run --runtime=nvidia --rm nvidia/cuda:9.1-runtime nvidia-smi

*目前15.8的nv驱动是390.68,只能支持cuda:9.1

Reply Favorite View the author
All Replies
avatar
a245404779
deepin
2018-12-26 08:26
#1
老哥牛逼!!
//修改/etc/nvidia-container-runtime/config.toml, 将ldconfig="@/sbin/ldconfig.real"注释掉
这个操作我做了

//增加/etc/ld.so.config.d/nvidia.conf
///usr/lib/x86_64-linux-gnu/nvidia
///usr/lib/x86_64-linux-gnu/nvidia/current
这几个操作没有做,但是测试通过了,问下上面两步分别有什么意义
Reply View the author
avatar
fengjl
deepin
2019-01-14 06:15
#2
standard_init_linux.go:190: exec user process caused "no such file or directory"
应该怎么处理,网络上提示dos2unix,不知道该怎么用
Reply View the author
avatar
fengjl
deepin
2019-01-14 06:16
#3
sudo nvidia-container-cli -k -d /dev/tty info

-- WARNING, the following logs are for debugging purposes only --

I0113 14:16:20.378423 6873 nvc.c:281] initializing library context (version=1.0.0, build=881c88e2e5bb682c9bb14e68bd165cfb64563bb1)
I0113 14:16:20.378502 6873 nvc.c:255] using root /
I0113 14:16:20.378506 6873 nvc.c:256] using ldcache /etc/ld.so.cache
I0113 14:16:20.378525 6873 nvc.c:257] using unprivileged user 65534:65534
I0113 14:16:20.379801 6874 nvc.c:191] loading kernel module nvidia
I0113 14:16:20.379998 6874 nvc.c:203] loading kernel module nvidia_uvm
I0113 14:16:20.380104 6874 nvc.c:211] loading kernel module nvidia_modeset
I0113 14:16:20.380353 6875 driver.c:133] starting driver service
I0113 14:16:20.403344 6873 nvc_info.c:433] requesting driver information with ''
I0113 14:16:20.403603 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libvdpau_nvidia.so.390.67
I0113 14:16:20.403705 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/libnvidia-tls.so.390.67
I0113 14:16:20.403734 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libnvidia-ptxjitcompiler.so.390.67
I0113 14:16:20.403788 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libnvidia-opencl.so.390.67
I0113 14:16:20.403837 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libnvidia-ml.so.390.67
I0113 14:16:20.403899 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libnvidia-ifr.so.390.67
I0113 14:16:20.403960 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/libnvidia-glsi.so.390.67
I0113 14:16:20.403983 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/libnvidia-glcore.so.390.67
I0113 14:16:20.404006 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libnvidia-fbc.so.390.67
I0113 14:16:20.404068 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/libnvidia-fatbinaryloader.so.390.67
I0113 14:16:20.404093 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libnvidia-encode.so.390.67
I0113 14:16:20.404151 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/libnvidia-eglcore.so.390.67
I0113 14:16:20.404173 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/libnvidia-compiler.so.390.67
I0113 14:16:20.404213 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libnvidia-cfg.so.390.67
I0113 14:16:20.404287 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libnvcuvid.so.390.67
I0113 14:16:20.404530 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libcuda.so.390.67
I0113 14:16:20.404750 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libGLX_nvidia.so.390.67
I0113 14:16:20.404803 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libGLESv2_nvidia.so.390.67
I0113 14:16:20.404899 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libGLESv1_CM_nvidia.so.390.67
I0113 14:16:20.405009 6873 nvc_info.c:147] selecting /usr/lib/x86_64-linux-gnu/nvidia/current/libEGL_nvidia.so.390.67
I0113 14:16:20.405098 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/libnvidia-tls.so.390.67
I0113 14:16:20.405140 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/nvidia/current/libnvidia-ptxjitcompiler.so.390.67
I0113 14:16:20.405182 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/nvidia/current/libnvidia-opencl.so.390.67
I0113 14:16:20.405223 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/nvidia/current/libnvidia-ml.so.390.67
I0113 14:16:20.405251 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/libnvidia-glsi.so.390.67
I0113 14:16:20.405278 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/libnvidia-glcore.so.390.67
I0113 14:16:20.405303 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/libnvidia-fatbinaryloader.so.390.67
I0113 14:16:20.405329 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/libnvidia-eglcore.so.390.67
I0113 14:16:20.405355 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/libnvidia-compiler.so.390.67
I0113 14:16:20.405408 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/nvidia/current/libnvidia-cfg.so.390.67
I0113 14:16:20.405471 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/nvidia/current/libcuda.so.390.67
I0113 14:16:20.405548 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/nvidia/current/libGLX_nvidia.so.390.67
I0113 14:16:20.405591 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/nvidia/current/libGLESv2_nvidia.so.390.67
I0113 14:16:20.405631 6873 nvc_info.c:147] selecting /usr/lib/i386-linux-gnu/nvidia/current/libEGL_nvidia.so.390.67
W0113 14:16:20.405651 6873 nvc_info.c:302] missing compat32 library libvdpau_nvidia.so
W0113 14:16:20.405657 6873 nvc_info.c:302] missing compat32 library libnvidia-encode.so
W0113 14:16:20.405663 6873 nvc_info.c:302] missing compat32 library libnvcuvid.so
W0113 14:16:20.405668 6873 nvc_info.c:302] missing compat32 library libnvidia-fbc.so
W0113 14:16:20.405674 6873 nvc_info.c:302] missing compat32 library libnvidia-ifr.so
W0113 14:16:20.405678 6873 nvc_info.c:302] missing compat32 library libGLESv1_CM_nvidia.so
I0113 14:16:20.405801 6873 nvc_info.c:228] selecting /usr/sbin/nvidia-cuda-mps-server
I0113 14:16:20.405830 6873 nvc_info.c:228] selecting /usr/lib/nvidia/current/nvidia-smi
I0113 14:16:20.405857 6873 nvc_info.c:228] selecting /usr/lib/nvidia/current/nvidia-debugdump
I0113 14:16:20.405872 6873 nvc_info.c:228] selecting /usr/bin/nvidia-persistenced
I0113 14:16:20.405887 6873 nvc_info.c:228] selecting /usr/bin/nvidia-cuda-mps-control
I0113 14:16:20.405907 6873 nvc_info.c:365] listing device /dev/nvidiactl
I0113 14:16:20.405911 6873 nvc_info.c:365] listing device /dev/nvidia-uvm
I0113 14:16:20.405916 6873 nvc_info.c:365] listing device /dev/nvidia-uvm-tools
I0113 14:16:20.405921 6873 nvc_info.c:365] listing device /dev/nvidia-modeset
W0113 14:16:20.405940 6873 nvc_info.c:273] missing ipc /var/run/nvidia-persistenced/socket
W0113 14:16:20.405952 6873 nvc_info.c:273] missing ipc /tmp/nvidia-mps
I0113 14:16:20.405958 6873 nvc_info.c:489] requesting device information with ''
I0113 14:16:20.411838 6873 nvc_info.c:519] listing device /dev/nvidia0 (GPU-23253bf8-4420-e4b0-cde3-d5cebac34933 at 00000000:01:00.0)
NVRM version:   390.67
CUDA version:   9.1

Device Index:   0
Device Minor:   0
Model:          GeForce GTX 1060
Brand:          GeForce
GPU UUID:       GPU-23253bf8-4420-e4b0-cde3-d5cebac34933
Bus Location:   00000000:01:00.0
Architecture:   6.1
I0113 14:16:20.411868 6873 nvc.c:318] shutting down library context
I0113 14:16:20.412134 6875 driver.c:192] terminating driver service
I0113 14:16:20.417225 6873 driver.c:233] driver service terminated successfully
Reply View the author
avatar
ljd3000
deepin
2019-01-14 17:39
#4
https://bbs.deepin.org/post/171918
老哥牛逼!!
//修改/etc/nvidia-container-runtime/config.toml, 将ldconfig="@/sbin/ldconfig.real"注释 ...

ldconfig是一个动态链接库管理命令,其目的为了让动态链接库为系统所共享
注释第一个,是因为deepin没有ldconfig.real
后面是把nvidia的相关动态库路径加到ldconfig的搜索路径中,不做可能也可以
Reply View the author
avatar
ljd3000
deepin
2019-01-14 17:46
#5
https://bbs.deepin.org/post/171918
standard_init_linux.go:190: exec user process caused "no such file or directory"
应该怎么处理,网络 ...

是不是nvidia-smi没有装amd64版本,如果是i386版本,就会报这个错误。当初研究了好久才发现这个问题。

$ apt search nvidia-smi

nvidia-smi/panda,now 390.67-3deepin amd64 [已安装]
  NVIDIA System Management Interface
Reply View the author
avatar
fengjl
deepin
2019-01-15 03:43
#6
https://bbs.deepin.org/post/171918
是不是nvidia-smi没有装amd64版本,如果是i386版本,就会报这个错误。当初研究了好久才发现这个问题。

$ ...

高手!果然这样就可以了
Reply View the author
avatar
markbrown
deepin
2019-06-23 17:21
#7
真NB
Reply View the author