问题描述

如题，起因是在阿里云GPU服务器上，使用原先正常运行的镜像生成了容器，但容器的显卡驱动出问题了，使用nvidia-smi命令会报错 NVIDIA-SMI has failed because it couldn’t communicate with the NVIDIA driver.

尝试使用官网.run文件重新安装显卡驱动会报错ERROR: Unable to load the kernel module ‘nvidia.ko’. This happens most frequently when this kernel module was built against the wrong or improperly configured kernel sources, with a version of gcc that differs from the one used to build the target kernel, or if a driver such as rivafb, nvidiafb, or nouveau is present and prevents the NVIDIA kernel module from obtaining ownership of the NVIDIA graphics device(s), or no NVIDIA GPU installed in this system is supported by this NVIDIA Linux graphics driver release.

按照报错信息，怀疑是内核版本或者gcc版本有误，更换了多个内核版本和gcc版本，使用了网上很多这两种保存相关的解决思路，都没能解决，一筹莫展。

放弃了原先的镜像，新建了空的容器，但是空的容器也会报NVIDIA-SMI has failed because it couldn’t communicate with the NVIDIA driver的错，并且空的容器居然也装不上显卡驱动，遂怀疑是容器本身的问题。

解决方案

发现可能是容器本身的设置有问题，设置为GPU计算时容器可正常安装驱动，但是设置为GPU计算可视化时就会报以上错误。
咨询阿里云，发现GPU计算可视化型需要提交工单获取特定的兼容驱动，GPU计算型才可以从官网下载驱动安装。通过提交工单获取特定的兼容驱动后，驱动可正常安装，问题解决。

反思

如果云服务器中空的容器连驱动都安装不好的话，就不要自己折腾了，大概率是容器本身哪里出问题了，咨询云服务商吧。

【已解决】nvidia-smi报错：NVIDIA-SMI has failed because it couldn’t communicate with the ... 阿里云GPU服务器相关推荐

nvidia-smi报错：NVIDIA-SMI has failed because it couldn‘t communicate with the NVIDIA driver 解决方案
nvidia-smi报错:NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure ...
nvidia-smi报错：NVIDIA-SMI has failed because it couldn‘t communicate with the NVIDIA driver 原因及避坑解决方案
nvidia-smi报错:NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver 原因及解决方案过了 ...
已解决（pip报错）WARNING: The repository located at mirrors .aliyun.com is not a trusted or secure host and
成功解决:WARNING: The repository located at mirrors .aliyun.com is not a trusted or secure host and is b ...
（已解决）STM32报错Error: L6218E: Undefined symbol assert_param (referred from misc.o).
今天在学习搭建固件库时遇到了一个问题,32在编译的时候会报Error: L6218E: Undefined symbol assert_param (referred from misc.o).的错误 ...
【已解决】chatgpt报错 access denied
报错原因 ip被封工具问题解决方法换节点或者换工具清空浏览器缓存,注意如果不行,就清空24小时或者更长时间的缓存,把时间线拉长. 补充:还是不行的话,就需要换工具,另外建议换成原生节点.
已解决：sts报错The fully qualified name of the bean's class, except if it serves only as..
我使用的ide为Spring Tools 4:在构建spring的项目进行bean的配置时,突然出现如下图所示错误: 翻译为bean类的命名必须是完全限定的,除非它是子类bean的父类描述. ---- ...
【已解决】DQN报错：NameError: name ‘glPushMatrix‘ is not defined
1.问题 pycharm在运行DQN平衡杆代码时报错:NameError: name 'glPushMatrix' is not defined. 画面只出现一个白色背景. 2.分析 pyglet版本 ...
已解决：Oracle报错 : input value not long enough for date format
先看一下报错信息翻译: 原因:是"输入参数"有问题,有可能是下面几种: 1.TIMESTAMP类型,用insert赋值时少了几位,毫秒数一般有13位,例如:157473693026 ...
解决argo workflow报错：MountVolume.SetUp failed for volume “docker-sock“ : hostPath type check failed
提交workflow时报错: MountVolume.SetUp failed for volume "docker-sock" : hostPath type check fai ...

【已解决】nvidia-smi报错：NVIDIA-SMI has failed because it couldn’t communicate with the ... 阿里云GPU服务器

问题描述

解决方案

反思

【已解决】nvidia-smi报错：NVIDIA-SMI has failed because it couldn’t communicate with the ... 阿里云GPU服务器相关推荐

最新文章

热门文章