tornado multi process pattern parsing

This paper mainly studies the related contents of tornado multi-process mode, which are as follows.

The boot method in the helloworld instance of the official document:

if __name__ == "__main__":

application. (8888) # listen is shortcut bind and start

 tornado.ioloop.IOLoop.instance().start()

And in listen, start tornado as a single-process model.

So the way to start a multi-process model is to get rid of the listen method

http_server = tornado.httpserver.HTTPServer(application)

  http_server.bind(options.port, options.host)

  http_server.start(num_processes=0) # tornado Will be in accordance with the cpu Auditing to fork process 

  tornado.ioloop.IOLoop.instance().start()

One thing to note is to turn off the debug mode, otherwise:

[E 110521 11:26:53 httpserver:229] Cannot run in multiple processes: IOLoop instance has already been initialized. You cannot call IOLoop.instance() before calling start()

The reason is that ES35en.py has already initialized IOLoop before ES37en_server.start (), which is explained in ES41en.py class HTTPServer() and def start().

conclusion

That's the end of this article on tornado multiprocess schema parsing, and I hope it's helpful. Interested friends can continue to refer to other related topics in this site, if there is any deficiency, welcome to comment out. Thank you for your support!