I've been playing around with the python API for memcached. Does anyone know if the client is intended to be used in a multithreaded environment (like for multiple web requests). Or is it more standard to 1) create a client for each request or 2) create a pool of clients. Thanks for any feedback. -Dennis