多显卡机器减少显卡后,docker无法启动对应gpu容器,提示:
Error response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: error during container init: error running hook #0: error running hook: exit status 1, stdout: , stderr: Auto-detected mode as 'legacy'
nvidia-container-cli: device error: 3: unknown device: unknown
Error: failed to start containers: ds
解决办法:
1、systemctl stop docker
2、编辑对应容器的docker配置文件hostconfig.json
删除DeviceIDs对应的KV值对,保留Capabilities,如下:

3、systemctl restart docker