Question

使用python，我并行启动两个子进程。一个是HTTP服务器，而另一个是另一个程序的执行（CustomSimpleHTTPServer.py，这是一个由selenium IDE插件生成的python脚本，用于打开firefox，导航到一个网站并进行一些交互）。另一方面，我想在第二个子进程完成执行时停止执行第一个子进程（HTTP Server）。

我的代码的逻辑是selenium脚本将打开一个网站。该网站将自动对我的HTTP服务器进行几次GET调用。在selenium脚本执行完毕后，应该关闭HTTP Server，以便它可以将所有捕获的请求记录在文件中。

这是我的主要Python代码：

class Myclass(object):

HTTPSERVERPROCESS = ""

    def startHTTPServer(self):
        print "********HTTP Server started*********"
        try:
            self.HTTPSERVERPROCESS=subprocess.Popen('python CustomSimpleHTTPServer.py', \
                            shell=True, stdout=subprocess.PIPE,stderr=subprocess.STDOUT)
        except Exception as e:
            print "Exception captured while starting HTTP Server process: %s\n" % e

    def startNavigatingFromBrowser(self):
        print "********Opening firefox to start navigation*********"
        try:
            process=subprocess.Popen('python navigationScript.py', \
                            shell=True, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
            process.communicate()
            process.wait()
        except Exception as e:
            print "Exception captured starting Browser Navigation process : %s\n" % e
        try:
            if process.returncode==0:
                print "HTTPSERVEPROCESS value: %s" % self.HTTPSERVERPROCESS.returncode
                print self.HTTPSERVERPROCESS
                #self.HTTPSERVERPROCESS.kill()
                #self.HTTPSERVERPROCESS.terminate()
                #self.kill(self.HTTPSERVERPROCESS.pid)
        except Exception as e:
            print "Exception captured while killing HTTP Server process : %s\n" % e

    def kill(self,proc_pid):
        process = psutil.Process(proc_pid)
        for proc in process.get_children(recursive=True):
            proc.kill()
        process.kill()

    def startCapture(self):
        print "********Starting Parallel execution of Server initiation and firefox navigation script*********"
        t1 = threading.Thread(target=self.startHTTPServer())
        t2 = threading.Thread(target=self.startNavigatingFromBrowser())
        t1.start()
        t2.start()
        t2.join()

注意：通过调用startCapture（）

开始执行

以下是CustomSimpleHTTPServer.py的代码，它应该在终止时将捕获的请求写入logfile.txt：

import SimpleHTTPServer
import SocketServer

PORT = 5555

class MyHTTPHandler(SimpleHTTPServer.SimpleHTTPRequestHandler):
    log_file = open('logfile.txt', 'w')
    def log_message(self, format, *args):
        self.log_file.write("%s - - [%s] %s\n" %
                        (self.client_address[0],
                         self.log_date_time_string(),
                         format%args))

Handler = MyHTTPHandler
httpd = SocketServer.TCPServer(("", PORT), Handler)
httpd.serve_forever()

当我使用self.HTTPSERVERPROCESS.kill（）或self.HTTPSERVERPROCESS.terminate（）或os.kill（）时，我会在运行主Python代码时在终端中关注

********Starting Parallel execution of Server initiation and firefox navigation script*********
********SimpleHTTPServer started*********
********Opening firefox to start navigation*********
HTTPSERVEPROCESS value: <subprocess.Popen object at 0x1080f8410>
2459
Exception captured while killing HTTP Server process : [Errno 3] No such process


Process finished with exit code 0

当我使用self.kill（self.HTTPSERVERPROCESS.pid）时，我会在运行主要Python代码后在终端中关注

********Starting Parallel execution of Server initiation and firefox navigation script*********
********SimpleHTTPServer started*********
********Opening firefox to start navigation*********
HTTPSERVEPROCESS value: <subprocess.Popen object at 0x1080f8410>
2459
Exception captured while killing HTTP Server process : 'Process' object has no attribute 'get_children'


Process finished with exit code 0

以下3个中的任何一个都无法终止HTTPServer进程：

self.HTTPSERVERPROCESS.kill()
self.HTTPSERVERPROCESS.terminate()
self.kill(self.HTTPSERVERPROCESS.pid)

我知道CustomSimpleHTTPServer.py是正确的，因为当我单独运行它并手动浏览到网站，然后通过点击终端中的CTRL-c手动终止CustomSimpleHTTPServer.py脚本时，日志将填充在logfle.txt中。

我对代码进行了哪些更改，以使其正常运行并填充日志？

Answer 1

您应该使用os.kill()来表示流程：

import os
import signal
...
os.kill(the_pid, signal.SIGTERM)   # usually kills processes
os.kill(the_pid, signal.SIGKILL)   # should always kill a process

此外，如果您杀死父进程，它通常也会杀死子进程。

更新

我对Server程序进行了两处小改动：

添加对self.log_file.flush()的调用以确保日志条目被刷新到日志文件中。
覆盖allow_reuse_address，以便您可以重复使用终止服务器后不久，同一地址。（见this SO question）

档案Server：

#!/usr/bin/env python

import SimpleHTTPServer
import SocketServer
PORT = 5555

class MyServer(SocketServer.TCPServer):
    allow_reuse_address = True

class MyHTTPHandler(SimpleHTTPServer.SimpleHTTPRequestHandler):
    log_file = open('logfile.txt', 'w')
    def log_message(self, format, *args):
        self.log_file.write("%s - - [%s] %s\n" %
                        (self.client_address[0],
                         self.log_date_time_string(),
                         format%args))
        self.log_file.flush()

Handler = MyHTTPHandler
httpd = MyServer(("", PORT), Handler)
httpd.serve_forever()

这是一个简单的导航程序（文件Navigate）：

#!/usr/bin/env python

import requests
import time

URL = "http://localhost:5555/"

def main():
  for x in range(5):
    print "x =", x
    r = requests.get(URL + "asd")
    time.sleep(1)
  print "Quitting"

main()

以下是主要计划：

#!/usr/bin/env python

import os
import subprocess
import signal

def main():
  # start web server
  web = subprocess.Popen(["./Server"])
  print "web server pid:", web.pid

  # start navigator
  nav = subprocess.Popen(["./Navigate"])
  print "nav pid: ", nav.pid

  # wait for nav to exit
  nav.wait()
  print "done waiting for nav"
  print "killing web server"
  os.kill(web.pid, signal.SIGTERM )
  web.wait()
  print "server terminated"

main()

无法使用process.kill（）或process.terminate（）或os.kill（）或使用psutil

1 个答案: