Re: [Twisted-Python] intermittent traceback using adbapi with pyPgSQL

It sounds like you have a database command running in the threadpool after the reactor has already been shut down. Which suggests you just want to make sure reactor.stop() is called after you're certain there are no more database calls running (and you mustn't call reactor.stop() in a thread of course - reactor.callFromThread(reactor.stop) if you want to).