AWS Lambda中Python+Selenium Chrome启动失败问题求助

在AWS Lambda上运行基于Python+Selenium的函数，用于启动Chrome浏览器访问指定网站。初始几次调用完全正常，但后续调用时出现报错，尝试多种方法仍未解决。

报错信息：

Message: session not created: Chrome failed to start: exited normally.
(chrome not reachable)
(The process started from chrome location /opt/chrome/chrome is no longer running, so ChromeDriver is assuming that Chrome has crashed.)

现有配置 #

驱动创建代码 #

chrome_options = Options()

service = Service("/opt/chromedriver")
chrome_options.binary_location = '/opt/chrome/chrome'
chrome_options.add_argument("--no-sandbox") # No protection needed
chrome_options.add_argument("--headless=new") # Hide the GUI
chrome_options.add_argument("--single-process") # Lambda only give us only one CPU
chrome_options.add_argument("--disable-dev-shm-usage") 
chrome_options.add_argument("--disable-extensions"); # disabling extensions
chrome_options.add_argument("--disable-gpu"); # applicable to windows os only
chrome_options.add_argument("--disable-infobars"); # disabling infobars
chrome_options.add_argument('--remote-debugging-port=9222')    

web_driver = webdriver.Chrome(options=chrome_options, service=service)

Dockerfile配置 #

FROM public.ecr.aws/lambda/python@sha256:a8770bd841c9caa98557ef7dc590bbd2707eb5ecab13bc8edbc81bf4961dca61 AS build
RUN yum install -y unzip && \
    curl -Lo "/tmp/chromedriver-linux64.zip" "https://storage.googleapis.com/chrome-for-testing-public/122.0.6261.69/linux64/chromedriver-linux64.zip" && \
    curl -Lo "/tmp/chrome-linux64.zip" "https://storage.googleapis.com/chrome-for-testing-public/122.0.6261.69/linux64/chrome-linux64.zip" && \
    unzip /tmp/chromedriver-linux64.zip -d /opt/ && \
    unzip /tmp/chrome-linux64.zip -d /opt/

FROM public.ecr.aws/lambda/python@sha256:a8770bd841c9caa98557ef7dc590bbd2707eb5ecab13bc8edbc81bf4961dca61
RUN yum install atk cups-libs gtk3 libXcomposite alsa-lib \
    libXcursor libXdamage libXext libXi libXrandr libXScrnSaver \
    libXtst pango at-spi2-atk libXt xorg-x11-server-Xvfb \
    xorg-x11-xauth dbus-glib dbus-glib-devel -y

# Copying requirements.txt and installing them
COPY requirements.txt ./
RUN pip install -r requirements.txt

# Copy the chrome binary and driver
COPY --from=build /opt/chrome-linux64 /opt/chrome
COPY --from=build /opt/chromedriver-linux64 /opt/

#Copy the python program file
COPY selenium_python_lambda.py ./

#Run the handler
CMD [ "selenium_python_lambda.lambda_handler" ]

requirements.txt #

selenium==4.10.0
unidecode==1.3.4
bs4==0.0.1

针对Lambda执行环境的特性，从以下几个方向调整：

• 移除远程调试端口参数
  删除chrome_options.add_argument('--remote-debugging-port=9222')。Lambda执行容器会被复用，第一次调用占用9222端口后，后续调用无法再次绑定该端口，直接导致Chrome启动失败。无头模式下无需该参数。

• 强制释放WebDriver资源
  在Lambda handler的逻辑末尾，必须调用web_driver.quit()而非web_driver.close()。quit()会彻底终止Chrome进程和驱动服务，避免残留进程占用资源影响后续调用。示例：

  def lambda_handler(event, context):
      try:
          # 业务逻辑代码
          web_driver.get("https://example.com")
          # ...
      finally:
          # 无论执行成功与否都释放资源
          web_driver.quit()
  

• 添加执行权限
  在Dockerfile中复制Chrome和驱动文件后，添加权限设置：

  COPY --from=build /opt/chrome-linux64 /opt/chrome
  COPY --from=build /opt/chromedriver-linux64 /opt/
  # 添加执行权限
  RUN chmod +x /opt/chrome/chrome /opt/chromedriver
  

  避免因文件无执行权限导致Chrome无法启动。

• 优化Chrome启动参数
  去掉可能引发稳定性问题的--single-process，同时添加适配Lambda环境的参数：

  # 移除--single-process
  chrome_options.add_argument("--disable-background-networking")
  chrome_options.add_argument("--disable-ipc-flooding-protection")
  chrome_options.add_argument("--disable-threaded-animation")
  chrome_options.add_argument("--disable-threaded-scrolling")
  

• 升级Selenium版本
  将requirements.txt中的Selenium版本升级至最新稳定版（如4.18.1），修复旧版本的兼容性问题：

  selenium==4.18.1
  unidecode==1.3.4
  bs4==0.0.1
  

内容的提问来源于stack exchange，提问作者S Dhas