How to build a thread-safe Request Handler with Python Tornado

Question

Does Tornado make any guarantees about how many requests are concurrently handled by a single RequestHandler? I'm thinking about managing session data like this: ``` class MyHandler(tornado.web.RequestHandler): def prepare(self): self.session = load_session_from_memcached def get(self): # work with self.session def on_finish(self): save_session_to_memcached(self.session) ``` But this is only going to work if `self.session` in `get()` is the same as `self.session` in `prepare()`. And that is only going to be true if a single RequestHandler processes at most one request concurrently. Is this how Tornado works?

Accepted Answer

`RequestHandler`s are instantiated per request. A single request handler instance will serve one and only one request in its lifetime

Edit:

Here's the link to the part of the documentation which confirms my statement above: http://www.tornadoweb.org/documentation/overview.html#overriding-requesthandler-methods

See Item 1: A new RequestHandler object is created on each request

How to build a thread-safe Request Handler with Python Tornado

Solution

Problem