通過 Crontab 使用 Selenium 執行 Python 腳本

Question

我有幾個在 Debian 服務器上使用 selenium webdriver 的 python 腳本。 如果我從終端（通常以 root 用戶）手動運行它們，一切正常，但是每次我嘗試通過 crontab 運行它們時，我都會遇到這樣的異常：

WebDriverException: Message: Can't load the profile. Profile Dir: /tmp/tmpQ4vStP If you specified a log_file in the FirefoxBinary constructor, check it for details.

試試這個腳本：

from selenium.webdriver.firefox.firefox_binary import FirefoxBinary
from pyvirtualdisplay import Display
from selenium import webdriver
import datetime
import logging

FIREFOX_PATH = '/usr/bin/firefox'

if __name__ == '__main__':
    cur_date = datetime.datetime.now().strftime('%Y-%m-%d')
    logging.basicConfig(filename="./logs/download_{0}.log".format(cur_date),
                        filemode='w',
                        level=logging.DEBUG,
                        format='%(asctime)s - %(levelname)s - %(message)s')
    try:
        display = Display(visible=0, size=(800, 600))
        display.start()
        print 'start'
        logging.info('start')
        binary = FirefoxBinary(FIREFOX_PATH,
                               log_file='/home/egor/dev/test/logs/firefox_binary_log.log')
        driver = webdriver.Firefox()
        driver.get("http://google.com")
        logging.info('title: ' + driver.title)
        driver.quit()
        display.stop()
    except:
        logging.exception('')
    logging.info('finish')
    print 'finish'

它的 crontab 命令：

0 13 * * * cd "/home/egor/dev/test" && python test.py

此腳本的日志文件如下所示：

2016-09-27 16:30:01,742 - DEBUG - param: "['Xvfb', '-help']" 
2016-09-27 16:30:01,743 - DEBUG - command: ['Xvfb', '-help']
2016-09-27 16:30:01,743 - DEBUG - joined command: Xvfb -help
2016-09-27 16:30:01,745 - DEBUG - process was started (pid=23042)
2016-09-27 16:30:01,747 - DEBUG - process has ended
2016-09-27 16:30:01,748 - DEBUG - return code=0
2016-09-27 16:30:01,748 - DEBUG - stdout=
2016-09-27 16:30:01,751 - DEBUG - param: "['Xvfb', '-br', '-nolisten', 'tcp', '-screen', '0', '800x600x24', ':1724']" 
2016-09-27 16:30:01,751 - DEBUG - command: ['Xvfb', '-br', '-nolisten', 'tcp', '-screen', '0', '800x600x24', ':1724']
2016-09-27 16:30:01,751 - DEBUG - joined command: Xvfb -br -nolisten tcp -screen 0 800x600x24 :1724
2016-09-27 16:30:01,753 - DEBUG - param: "['Xvfb', '-br', '-nolisten', 'tcp', '-screen', '0', '800x600x24', ':1725']" 
2016-09-27 16:30:01,753 - DEBUG - command: ['Xvfb', '-br', '-nolisten', 'tcp', '-screen', '0', '800x600x24', ':1725']
2016-09-27 16:30:01,753 - DEBUG - joined command: Xvfb -br -nolisten tcp -screen 0 800x600x24 :1725
2016-09-27 16:30:01,755 - DEBUG - process was started (pid=23043)
2016-09-27 16:30:01,755 - DEBUG - DISPLAY=:1725
2016-09-27 16:30:01,855 - INFO - start
2016-09-27 16:30:31,965 - ERROR - 
Traceback (most recent call last):
  File "test.py", line 31, in <module>
    driver = webdriver.Firefox()
  File "/usr/local/lib/python2.7/dist-packages/selenium/webdriver/firefox/webdriver.py", line 103, in __init__
    self.binary, timeout)
  File "/usr/local/lib/python2.7/dist-packages/selenium/webdriver/firefox/extension_connection.py", line 51, in __init__
    self.binary.launch_browser(self.profile, timeout=timeout)
  File "/usr/local/lib/python2.7/dist-packages/selenium/webdriver/firefox/firefox_binary.py", line 68, in launch_browser
    self._wait_until_connectable(timeout=timeout)
  File "/usr/local/lib/python2.7/dist-packages/selenium/webdriver/firefox/firefox_binary.py", line 106, in _wait_until_connectable
    % (self.profile.path))
WebDriverException: Message: Can't load the profile. Profile Dir: /tmp/tmpQ4vStP If you specified a log_file in the FirefoxBinary constructor, check it for details.

2016-09-27 16:30:31,966 - INFO - finish

我嘗試過的：

1. 確保腳本文件歸 root 所有
2. 使用導出顯示=:0; 或導出 DISPLAY=:99; 在 crontab 命令中
3. 將 crontab 中的 HOME 變量設置為運行 cronjob 的用戶的主目錄的路徑

我真的被這個問題困住了。

我在 Debian 7.7 上有 python 2.7.10、selenium 2.53.6 和 Xvbf 和 Firefox 47.0.1

Answer 1

問題與環境變量有關：cron 是由系統啟動的，對用戶環境一無所知。

因此，解決問題的方法是讓cron 運行一個shell 腳本，該腳本先設置所需的環境變量，然后再運行該腳本。 在我的情況下，我需要像這樣設置PATH變量： PATH=/root/anaconda3/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin此外，在某些情況下設置HOME或DISPLAY變量可能很有用。

Answer 2

嘗試使用硬編碼的 Firefox 二進制文件https://seleniumhq.github.io/selenium/docs/api/py/webdriver_firefox/selenium.webdriver.firefox.firefox_binary.html

selenium.webdriver.firefox.firefox_binary.FirefoxBinary("/your/binary/location/firefox")
driver = webdriver.Firefox(firefox_binary=binary)