标签：
最近工作中遇到一个问题，在集群上运行的任务有时候无法正常结束，或者无法正常启动。这会造成这批运行的任务无法正常结束运行，处于pending的状态，导致后面的任务无法正常启动。
该问题困扰我们项目已经有半年左右了，一直没有想到很好的解决办法。主要原因就是任务的状态只能在浏览器中看出，无法通过后台的日志或者数据库查询得到。在浏览器中，如果我们看到某个任务长时间没有运行时间和状态的变化，就可以把这个任务当做是“僵尸”任务，从而可以将该任务手动结束掉（kill）。
春节之后在网上看到一些有关爬虫的文章，里面提到过有一种爬虫就是模拟浏览器的行为（包括登录、点击等）去得到网页的数据，进而进行网页抓取，有用信息提取。于是我思考，我们项目的问题和浏览器的交互，只有几种情况，完全可以通过这种方式解决“僵尸”任务。经过一周左右的研究和一周断断续续的coding，终于将这个问题解决了，现在把解决问题的主要思路和关键技术难点写下来，希望一来可以加深自己的印象，二来可以帮助到需要的人。因为实现的任务比较单一，且实现过程比较仓促，code主要就是实现了一些功能，没有进行优化，也没有太参考什么编码规范，设计模式之类的。以后遇到更大的问题，再考虑这些吧。
技术要点：
（1）Python的package:selenium，用这个package，可以和浏览器进行交互，如打开某个浏览器（Chrome，FireFox等），登录需要验证的网站（输入用户名&密码），点击某个特定图标等等，下面是两个有关selenium的链接：
https://www.baidu.com/link?url=tTeJRPOMKX8noXyTa2YPgpaD6vVlGQ2-RVAfwRg4Yvm&wd=&eqid=acd0879a0043c2e9000000045741cd39
http://www.cnblogs.com/fnng/archive/2013/05/29/3106515.html
（2）selenium的PhantomJS，这是一个虚拟的浏览器，可以把它看成一个在后台运行的浏览器，用户看不到浏览器的页面，但其他的功能和普通浏览器基本一样，比如可以截图，点击某个图标，抓取网页信息等，之所以使用了这个用来模仿浏览器，是因为我们的server无法安装普通的浏览器，只能运行在终端模式下运行的程序；
http://phantomjs.org/
（3）xpath，这个是我编程中耗时最多的模块，主要原因有几个，一是元素定位有问题，网站是一秒钟刷新一次，上一秒获取到的元素下一秒就找不到了；二是相似元素太多，层级关系太复杂，用一般的相对路径去寻找，有可能找到一些不想要的元素，所以就造成了寻找元素过程的费时费力。下面是两个有关xpath的介绍，比较实用，特别是在网页爬虫方面（后面我还要专门介绍爬虫）：
http://www.cnblogs.com/fdszlzl/archive/2009/06/02/1494836.html
http://www.ruanyifeng.com/blog/2009/07/xpath_path_expressions.html
以下是核心code，因为项目隐私的原因，把一些敏感的内容用*******代替。如果有什么问题，可以给我留言。
‘‘‘
command:
python tools/webScrap/KillJobs.py -url=******** -screenShotPath=******
‘‘‘
from selenium import webdriver
from selenium.common.exceptions import NoSuchElementException
from selenium.webdriver.common.keys import Keys
import re
import time
import argparse
import sys
import os
mailReceiver = [
*********,
********
]
ZOMBIE_JOB_LIST = {"list1": [], "list2": [], "list3":[]}
def getMailReceiver():
receiver = ‘ ‘
for recv in mailReceiver:
receiver = receiver + recv + ‘ ‘
return receiver
def kill_zombie_jobs(screenShotPath, url):
# browser = webdriver.PhantomJS() # Get local session of PhantomJS
browser = webdriver.Firefox() # Get local session of Firefox
browser.set_window_size(1600, 1000)
targetUrl = "http://%s/#JOBS" %url
print "url: ", targetUrl
job_to_be_kill_indicate = 0
browser.get(targetUrl) # Load page
userName = browser.find_elements_by_class_name("gwt-TextBox")
password = browser.find_elements_by_class_name("gwt-PasswordTextBox")
submitButton = browser.find_elements_by_class_name("gwt-Button")
if len(userName) == 0 or len(password) == 0 or len(submitButton) == 0:
print "error in open url: %s" %targetUrl
browser.quit()
return
userName[0].send_keys("******")
password[0].send_keys("*******")
time.sleep(1)
submitButton[0].click()
time.sleep(4)
sceen_shot_name = screenShotPath + "/Before_kill_jobs_screen_shot.png"
browser.save_screenshot(sceen_shot_name)
# get the Job_Name
# job_name_pattern = "/html/body/div[2]/div[2]/div/div[4]/div/div[3]/div/div[4]/div/div[2]/div/div[2]/div/div/div/div[3]/table[2]/tbody/tr[1]/td/fieldset/table/tbody/tr/td/table/tbody/tr[1]/td/table/tbody[1]/tr[1]/td[1]/div"
# jobs_name_pattern_0 = "//body/div[2]/div[2]/div/div[4]/div/div[3]/div/div[4]/div/div[2]/div/div[2]/div/div/div/div[3]/table[2]/tbody/tr[1]/td/fieldset/table/tbody/tr/td/table/tbody/tr/td/table/tbody/tr/td[@class=‘GJOFO-MDOC GJOFO-MDAD GJOFO-MDBD‘ or @class=‘GJOFO-MDOC GJOFO-MDAE GJOFO-MDBD‘]"
jobs_name_pattern_0 = "//body/div[2]/div[2]/div/div[4]/div/div[3]/div/div[4]/div/div[2]/div/div[2]/div/div/div/div[3]/table[2]/tbody/tr[1]/td/fieldset/table/tbody/tr/td/table/tbody/tr/td/table/tbody/tr/td[1]"
jobs_name_pattern = "//body/div[2]/div[2]/div/div[4]/div/div[3]/div/div[4]/div/div[2]/div/div[2]/div/div/div/div[3]/table[2]/tbody/tr[1]/td/fieldset/table/tbody/tr/td/table/tbody/tr/td/table/tbody/tr[Order]/td[1]"
jobs_kill_pattern = "//body/div[2]/div[2]/div/div[4]/div/div[3]/div/div[4]/div/div[2]/div/div[2]/div/div/div/div[3]/table[2]/tbody/tr[1]/td/fieldset/table/tbody/tr/td/table/tbody/tr/td/table/tbody/tr[Order]/td[3]"
jobs_duration_pattern = "//body/div[2]/div[2]/div/div[4]/div/div[3]/div/div[4]/div/div[2]/div/div[2]/div/div/div/div[3]/table[2]/tbody/tr[1]/td/fieldset/table/tbody/tr/td/table/tbody/tr/td/table/tbody/tr[Order]/td[5]"
for i in range(1, 4):
tmp_list = "list"+str(i)
job_name_elements_list = browser.find_elements_by_xpath(jobs_name_pattern_0)
job_length = len(job_name_elements_list)
for index in range(1, job_length+1):
job_name_pattern = jobs_name_pattern.replace("Order", str(index))
job_duration_pattern = jobs_duration_pattern.replace("Order", str(index))
job_name = get_element_name(browser, job_name_pattern)
job_duration_time = durationTime(get_element_name(browser, job_duration_pattern))
if len(job_name) > 10 and durationTime == 0:
ZOMBIE_JOB_LIST[tmp_list].append(job_name)
time.sleep(60)
print "zombie job list: ", ZOMBIE_JOB_LIST
job_name_elements_list = browser.find_elements_by_xpath(jobs_name_pattern_0)
job_length = len(job_name_elements_list)
for index in range(1, job_length+1):
# print "index:", index
job_name_pattern = jobs_name_pattern.replace("Order", str(index))
job_kill_pattern = jobs_kill_pattern.replace("Order", str(index))
job_name = get_element_name(browser, job_name_pattern)
if job_name in ZOMBIE_JOB_LIST["list1"] and job_name in ZOMBIE_JOB_LIST["list2"] and job_name in ZOMBIE_JOB_LIST["list3"]:
job_to_be_kill_indicate = 1
print "this job is should be killed: ", job_name
kill_button_element = browser.find_element_by_xpath(job_kill_pattern)
kill_button_element.click()
time.sleep(1)
confirm_kill_button_pattern = "//tbody/tr/td[1]/button[@class=‘gwt-Button‘]"
confirm_kill_button_element = browser.find_element_by_xpath(confirm_kill_button_pattern)
confirm_kill_button_element.click()
time.sleep(2)
time.sleep(2)
sceen_shot_name = screenShotPath + "/After_kill_jobs_screen_shot.png"
browser.save_screenshot(sceen_shot_name)
browser.quit()
return job_to_be_kill_indicate
def get_element_name(browser, element_pattern):
element_name = ""
try:
element = browser.find_element_by_xpath(element_pattern)
element_name = element.text
except Exception, e:
print "element not exist any more!!!!!"
element_name = ""
return element_name
def durationTime(timeStr):
if timeStr is None or timeStr == "":
return 0
if re.match(r"\d{2}:\d{2}:\d{2}", timeStr) is None:
return 0
timeSec = int(timeStr[0:2]) * 3600 + int(timeStr[3:5]) * 60 + int(timeStr[6:8])
return timeSec
def send_kill_jobs_mail(mailer, screenShotPath, url, indicator):
# jobs screen before and after kill
mailTitle = "Jobs_on_%s_Hanging" %url
screenShotFile1 = screenShotPath + "/Before_kill_jobs_screen_shot.png"
screenShotFile2 = screenShotPath + "/After_kill_jobs_screen_shot.png"
logFile = screenShotPath + "/nodes_hanging.log"
command = ‘mail -a ‘ + screenShotFile1 + ‘ -a ‘ + screenShotFile2 + ‘ -s ‘ + mailTitle + mailer + ‘ < ‘ + logFile
print "command: ", command
os.system(command)
return 0
def monitor():
# kill exist PhantomJS
command = "killall phantomjs"
print "kill all existing phantomjs: ", command
os.system(command)
parser = argparse.ArgumentParser()
parser.add_argument(‘-url‘, action=‘store‘, dest=‘url‘, help=‘data url‘, required=True)
parser.add_argument(‘-screenShotPath‘, action=‘store‘, dest=‘screenShotPath‘, help=‘the screen shot path‘, required=True)
results = parser.parse_args()
print ‘DataRush URL = ‘, results.url
url = results.url
print ‘Screen Shot Path = ‘, results.screenShotPath
screenShotPath = results.screenShotPath
mailer = getMailReceiver()
print "START: Monitor DataRush Starting.............................."
job_killed_inicate = kill_zombie_jobs(screenShotPath, url)
if job_killed_inicate == 1:
print "zombie jobs has been killed!!!!!!!"
send_kill_jobs_mail(mailer, screenShotPath, url, 1)
else:
pass
print "End: Monitor Finished....................................."
if __name__ == ‘__main__‘:
monitor()
标签：
原文地址：http://www.cnblogs.com/zhangchao3322218/p/5518379.html
