Question

我有一个简单的多进程池脚本，使用map来获取有序的结果，就像顺序执行一样。

注意 - 请忽略＆＃34;（1-3A-D）＆＃34;。它们不是输出的一部分，仅用于说明目的。假设它是序列多处理执行。

import base64
import getpass
import os

from ssh.ssh_module import SSH
from multiprocessing import Pool

username = ''
password = ''
cmd = 'uptime'

def runcommand (server):

    print ('Proccess ID: %d' % (os.getpid ()))
    ssh = SSH (server, username, password)
    if ssh.connect ():
        print (ssh.command (cmd))

    ssh.close ()


if __name__ == '__main__':

    username = input ('Username: ')
    # Encode password for SSH
    password = getpass.getpass ('Password: ')
    password = base64.b64encode (password.encode('utf-8'))

    servers = ['192.168.100.1', '192.168.100.2', '192.168.100.3']
    with Pool (processes = 2) as pool:
        pool.map (runcommand, servers)

输出：

(virtual) [user@centos python]$ ./multiprocess.py 
Username: spongebob
Password: 
(1A) Proccess ID: 24528
(1B) SSH into 192.168.100.1
(2A) Proccess ID: 24529
(2B) SSH into 192.168.100.2
(1C) Successfully logged in!
     00:15:24 up 4 days,  8:16,  0 users,  load average: 0.12, 0.10, 0.20

(1D) SSH session closed.

(3A) Proccess ID: 24529
(3B) SSH into 192.168.100.3
(2C) Successfully logged in!
     00:15:26 up 7 days,  6:17,  0 users,  load average: 0.10, 0.11, 0.09

(2D) SSH session closed.

(3C) Successfully logged in!
     00:15:47 up 5 days,  6:41,  0 users,  load average: 0.14, 0.10, 0.17

(3D) SSH session closed.

正如您所看到的，结果是随机返回的，我认为使用＆＃39; pool.map＆＃39;应该订购。首先执行IP /系统的顺序并不重要，但输出必须按正确的顺序排列才有意义。有关如何修复它的任何想法，以便我可以得到下面的欲望输出？

(1A) Proccess ID: 24528
(1B) SSH into 192.168.100.1
(1C) Successfully logged in!
     00:15:24 up 4 days,  8:16,  0 users,  load average: 0.12, 0.10, 0.20

(1D) SSH session closed.

(2A) Proccess ID: 24529
(2B) SSH into 192.168.100.2
(2C) Successfully logged in!
     00:15:26 up 7 days,  6:17,  0 users,  load average: 0.10, 0.11, 0.09

(2D) SSH session closed.

(3A) Proccess ID: 24529
(3B) SSH into 192.168.100.3
(3C) Successfully logged in!
     00:15:47 up 5 days,  6:41,  0 users,  load average: 0.14, 0.10, 0.17

(3D) SSH session closed.

Answer 1

Pool.map以与计划时相同的顺序返回结果，但不确保它们以相同的顺序执行。

顺序执行会破坏使用multiprocessing的目的。如果您需要按顺序执行某些操作，请使用map。

Answer 2

我设法通过删除SSH类中的消息并将其移动到调用者来实现所需的输出。这是使用.join（）来构造消息序列的问题。它不是最漂亮但现在已经订购了。

import base64
import getpass
import os

from cet.ssh_module import SSH
from multiprocessing import Pool

username = ''
password = ''
cmd = 'uptime'


def runcommand (server):

    print ('Proccess ID: %d' % (os.getpid ()))
    ssh = SSH (server, username, password)
    s1 = 'SSH into %s' % server
    if ssh.connect ():
        s2 = 'Successfully logged in!'
        s3 = ssh.command (cmd)

        ssh.close ()
        s4 = 'SSH session closed.\n'

        output = '\n'.join ([s1, s2, s3, s4])

        return output
    else:
        pass


if __name__ == '__main__':

    servers = ['192.168.100.1', '192.168.100.2', '192.168.100.3']

    username = input ('Username: ')
    # Encode password for SSH
    password = getpass.getpass ('Password: ')
    password = base64.b64encode (password.encode('utf-8'))

    with Pool (processes = 4) as pool:
        results = pool.map (runcommand, servers)

    for i in results:
        print (i)

<强>输出

(virtual) [user@centos python]$ ./multiprocess.py 
Username: spongebob
Password: 
Proccess ID: 40764
Proccess ID: 40763
Proccess ID: 40766
SSH into 192.168.100.1
Successfully logged in!
 17:25:33 up 8 days, 23:51,  0 users,  load average: 3.04, 2.05, 1.43

SSH session closed.

SSH into 192.168.100.2
Successfully logged in!
 17:25:17 up 2 days,  1:26,  0 users,  load average: 0.44, 0.42, 0.72

SSH session closed.

SSH into 192.168.100.3
Successfully logged in!
 17:25:15 up 8 days, 23:27,  0 users,  load average: 2.37, 2.35, 2.03

SSH session closed.

Python多处理映射返回无序结果

2 个答案: