Skip to content

README_chs.md

Latest commit

 

History

History
71 lines (66 loc) · 3.1 KB

README_chs.md

File metadata and controls

71 lines (66 loc) · 3.1 KB
Raw

Python 远程开发

介绍如何在windows系统上配置python的远程开发，IDE为pycharm，或者使用vscode作为编辑器。 [中文版][英文版]

环境

Platform

  • Windows 10/11
  • Pycharm 2023.1 专业版.
  • Vscode

Remote host

  • Login host, (called login0)
    • Port 22 is open for the user (only)
    • Support ssh connection with port 22 from WAN.
  • computer host, (called gpu0)
    • Conda environment, python
    • Port 22 is open for the user (only)
    • support ssh connection from WAN
      • May not support being connected from WAN, similar to the container.
  • the docker's container, which runs on the computer host (called container)
    • Conda environment, python
    • It cannot be connected by the user from the outside of the computer host (or LAN).

分析与解决

  • 首先, 需要配置本地的ssh默认配置文件 ../.ssh/config. 建议使用默认目录保存公钥和私钥，以下均为默认路径下的示例.

      C:\Users\account_name\.ssh

  #-------example--------
  C:\Users\ironieser\.ssh
  #----------------------

    参照: https://zhuanlan.zhihu.com/p/28423720

    Host login01
    HostName 10.15.15.15
    User your_name
    Port your_port
#-------example--------
Host login01
    HostName 10.15.15.15
    User ironieser
    Port 223
#----------------------
    • 需要将公钥文件id_rsa.pub 上传到登陆节点，计算节点和容器内（打包镜像保存）。

  • 登陆节点/堡垒机

    • Support straight connection

  • 计算节点

    1. 支持局域网或公网访问
    2. 仅支持局域网访问
    • 需要端口转发或多级跳转，下面为多级跳转示例
      • ssh -J login0 gpu0
      • ssh -J login0 gpu0

  • Docker

    • 运行容器，并同时开启端口转发，如示例，将端口12555转发到端口2555，容器内配置ssh端口为2555，实现通过连接12555，进而连接容器。

      docker run -it -v /public/homes/<usr_name>/project:/workdir  -w /workdir -p <contrainer_ssh_port_forward>:<container_ssh_port_listenning> --name=<your_container_name> <image_name>:<image_tag> <your shell>

#-------example--------
podman run -it -v /public/homes/ironieser/project:/workdir  -w /workdir -p 12555:2555 --name=talklip1 localhost/ironieser:zsh-cuda11.3-cudnn8-devel-ubuntu20.04 /bin/zsh
#----------------------

    • 在本地cmd/shell中开启端口转发

      ssh -N -f -L <local_host port>:<computer_host>:<container port> -p <computer_host port> name@<computer_host> -o TCPKeepAlive=yes
ssh -N -f -L 6001:192.168.4.129:12555 -p 22 ironieser@192.168.4.129

    • 在pycharm中添加ssh配置
      host IP: 127.0.0.1, port: 6001, name: root <container user name>

    • 现在，可以从公网直接访问计算节点或是容器了，并且支持远程调试