如何解决Docker容器内mpirun调用Python脚本失败的问题？

我尝试用以下Python脚本通过Docker SDK在容器内执行runScript_v2.py：

import docker

# Initialize the Docker client
client = docker.from_env()

# Define Docker run options
docker_run_options = {
    'name': 'container-name',
    'user': 'dafoamuser',
    'mounts': [
        {
            'type': 'bind',
            'source': 'D:/Dafoam/oneraM6',
            'target': '/home/dafoamuser/mount'
        }
    ],
    'working_dir': '/home/dafoamuser/mount',
    'detach': True,
    'tty': True,
    'stdin_open': True,
    'command': 'bash'
}

# Run the Docker container with specified options
container = client.containers.run(
    'dafoamimage:latest',
    **docker_run_options
)

# Execute the mpirun command inside the container
mpirun_command = 'mpirun -np 8 python runScript_v2.py 2>&1 | tee logOpt.txt'
exec_cmd = container.exec_run(mpirun_command, tty=True, privileged=True)
print(f"Command: {mpirun_command}\nOutput:\n{exec_cmd.output.decode('utf-8')}")

# Stop and remove the container (if needed)
container.stop()
container.remove()

运行后出现如下错误：

C:\Users\yesbo\.conda\envs\py39env\lib\site-packages\paramiko\transport.py:219: CryptographyDeprecationWarning: Blowfish has been deprecated
  "class": algorithms.Blowfish,
Command: mpirun -np 8 python runScript_v2.py 2>&1 | tee logOpt.txt
Output:
--------------------------------------------------------------------------
mpirun was unable to find the specified executable file, and therefore
did not launch the job.  This error was first reported for process
rank 0; it may have occurred for other processes as well.

NOTE: A common cause for this error is misspelling a mpirun command
      line parameter option (remember that mpirun interprets the first
      unrecognized command line token as the executable).

Node:       c537205732b5
Executable: python
--------------------------------------------------------------------------
8 total processes failed to start

直接在容器终端运行该mpirun命令可正常执行，求解决办法。

• 使用python的绝对路径
  先在容器终端执行which python获取python的绝对路径（比如/usr/bin/python），然后修改mpirun命令：

  mpirun_command = 'mpirun -np 8 /usr/bin/python runScript_v2.py 2>&1 | tee logOpt.txt'
  

  原因：exec_run执行命令时的环境变量PATH可能和终端登录时不同，导致无法找到python可执行文件。

• 通过shell执行命令
  exec_run默认直接执行命令，不经过shell，管道|和重定向2>&1无法生效，同时环境变量也不完整。修改命令为通过bash执行：

  mpirun_command = '/bin/bash -c "mpirun -np 8 python runScript_v2.py 2>&1 | tee logOpt.txt"'
  exec_cmd = container.exec_run(mpirun_command, tty=True, privileged=True)
  

• 显式指定环境变量
  在exec_run时传入完整的PATH环境变量，确保能找到python：

  exec_cmd = container.exec_run(
      mpirun_command, 
      tty=True, 
      privileged=True,
      environment={"PATH": "/usr/local/bin:/usr/bin:/bin:/home/dafoamuser/.local/bin"}
  )
  

  路径可根据容器内实际环境调整。

内容的提问来源于stack exchange，提问作者Mohsin Bashir