[rabbitmq-discuss] [Bug] Memory leak in 3.1.3 (Ubuntu)
louis.bouchard at canonical.com
Fri Jul 5 13:38:21 BST 2013
Le 05/07/2013 14:24, Tim Watson a écrit :
> Well there's a good reason why that leaks I think. When is the test case ever actually getting into the ./queue_receive bit!? Isn't that like doing `while true; do echo "hi"; done; echo "done"` - you never arrive at "done" do you.
> Even if you were calling the receive script in the loop, that's creating a new consumer and not removing it, so the memory use for all those connections will just keep going up since they're never being closed.
> Have I missed something? It really looks like this is just a badly formed test case to me.
Indeed, now that I look at it (I'm not the author of the test case, just
working on that bug) the syntax does look somewhat *doubtful*.
Let me run another set of tests to verify all this before bringing some
more noise to this list.
If anything gets confirmed, I will have rabbitmqctl status, rabbitmqctl
thanks for your help,
Backline Support Analyst
Ubuntu support: http://canonical.com/support
More information about the rabbitmq-discuss