Tornado timeouts and server failured

Question

I am working now on real-time game based on tornado, tornado-sockjs. There are a lot of different timeout strategies in our game application: TIMEOUT_GAME_IF_NOBODY, TIMEOUT_GAME_IF_SERVER_OFF. These timeouts has callbacks that can work with storage directly (update, insert, and so on). The question is what is the right way to organize timeout strategy into a module ?! How can we re-execute callbacks in case of server failure ? Imagine that three timeouts are hanging, and suddenly server that handles these timeouts, crashed. It means that some information was not updated.

Answer 1

You should send all importent data through queue with delivery confirm. So if you server will crash all data will come to it from queue. Try to use rabbitmq.