Selenium 没有关闭 docker 中的浏览器进程

Question

I have a docker container:

    ...
    web:
     build:
      context: .
      dockerfile: Dockerfile
     command: python /code/manage.py migrate --noinput
     command: python /code/manage.py runserver 0.0.0.0:8000
     volumes:
      - .:/code
     ports:
      - 80:8000
     environment:
      - SECRET_KEY=changemeinprod
     depends_on:
      - db
      - redis
     ...

Selenium scrape runs in this container (django project in docker):

     ...
     chrome_options = Options()
     chrome_options.add_argument("--headless")
     chrome_options.add_argument('--no-sandbox')
     chrome_options.add_argument('--disable-dev-shm-usage')
     prefs = {"profile.managed_default_content_settings.images": 2}
     chrome_options.add_experimental_option("prefs", prefs)

     driver = webdriver.Chrome(executable_path=chrome_driver_path, 
     options=chrome_options)
     driver.get(link)
     ...
     ...
     pid = driver.service.process.pid

     driver.quit()
     try:
         os.kill(int(pid), signal.SIGTERM)
         print("Killed chrome using process")
     except ProcessLookupError as e:
         pass

Despite I close driver with command driver.quit() , chrome process remains running. Every executing of this script keeps chrome process and makes next execution slower, because of multiple chrome processes running. How can I achieve closing chrome inside docker? Even If I need to restart container where django project running, can you tell me how to do it in python script (docker container restarts itself with python script).

Answer 1

I suggest you use Zalenium. It is a Selenium Grid wrapper which facilitates auto scaling/descaling. It will make your life much easier.

https://opensource.zalando.com/zalenium/

Answer 2

Try using driver.close() instead of driver.quit() . Also, try adding timeout to your call out... It may just be the link taking too long to download.