Question

我通过Elastic Beanstalk在AWS上运行了一个CherryPy应用程序，其中一项工作是检查Gmail收件箱中的新邮件，如下所示：

编辑：

# snippet of the code responsible for starting the poller in question
# application.py
def __init__(self):
    ...

    # register callbacks with monitor (token access check, history feed poller)
    credentials_checker = Monitor(cherrypy.engine, self.check_access_token, frequency=30)
    mail_checker = Monitor(cherrypy.engine, self.mail_processor.safe_poll_history_feed, frequency=10)
    credentials_checker.subscribe()
    mail_checker.subscribe()

# mail_checker.py
def safe_poll_history_feed(self):
    try:
        self.poll_history_feed()
    except:
        # no restart needed
        cherrypy.log("Exception in poll_history_feed()!", traceback=True)

def poll_history_feed(self):
    new_messages = self.mail_service.users().messages().list(userId="me", labelIds=["INBOX", "UNREAD"]).execute()

    result_size = int(new_messages["resultSizeEstimate"])
    cherrypy.log("Checking for new messages...")

    if result_size > 0:
        # process message
        cherrypy.log("number of new messages: %s" % result_size)
        cherrypy.log("New messages! Processing...")
        for x in new_messages["messages"]:
            self.process_email(x["id"])
    else:
        cherrypy.log("No new messages.")
        return

但问题是，似乎负责进行实际检查的线程似乎被某些东西杀死了，我无法弄清楚是什么。下面的日志片段：

[14/Mar/2016:15:32:55] Checking for new messages...
[14/Mar/2016:15:32:55] No new messages.
[14/Mar/2016:15:33:05] Checking for new messages...
[14/Mar/2016:15:33:05] No new messages.
[14/Mar/2016:15:33:15] Checking for new messages...
[14/Mar/2016:15:33:15] No new messages.
[14/Mar/2016:15:33:17] Checking access token.
[14/Mar/2016:15:33:17] Access token still good. Continuing...
[14/Mar/2016:15:33:26] Checking for new messages...
[14/Mar/2016:15:33:26] No new messages.
[14/Mar/2016:15:33:32] ENGINE Keyboard Interrupt: shutting down bus
[14/Mar/2016:15:33:32] ENGINE Bus STOPPING
[14/Mar/2016:15:33:32] ENGINE HTTP Server cherrypy._cpwsgi_server.CPWSGIServer(('127.0.0.1', 8080)) shut down
[14/Mar/2016:15:33:32] ENGINE Stopped thread 'Monitor'.
[14/Mar/2016:15:33:32] ENGINE Stopped thread 'Autoreloader'.
[14/Mar/2016:15:33:32] ENGINE Stopped thread 'Monitor'.
[14/Mar/2016:15:33:32] ENGINE Stopped thread '_TimeoutMonitor'.
[14/Mar/2016:15:33:32] ENGINE Bus STOPPED
[14/Mar/2016:15:33:32] ENGINE Bus EXITING
[14/Mar/2016:15:33:32] ENGINE Bus EXITED
[14/Mar/2016:15:33:32] ENGINE Waiting for child threads to terminate...
# note: the server seems to be responding to outside requests judging by this line and the ones following it.
[16/Mar/2016:04:48:41] Sent internal chat notification to user@userdomain.com, subject [MENTIONED] Client Portal 
...SNIP

该应用程序的其余部分似乎保持不变，因为它响应请求就好了，但它完全停止检查新邮件。重新启动应用服务器只能暂时修复它（不到10分钟左右）。我已经检查过，我可以考虑检查一下，但是在这个过程被杀的时候我还没有看到其他任何事情发生。

自从问题首次发布以来，我已经增加了实例类型的大小，只是为了看看是否能解决这个问题（它没有做到）。我想我可能在某些时候搞砸了日志，因为我不再在日志中看到来自CherryPy的ENGINE消息，所以我现在正在努力解决这个问题。

可能导致这种情况的原因，我该如何解决？我猜测它可能检查邮件的频率有点太快了？

Cherrypy公共汽车停止了一些未知的原因

0 个答案: