[rabbitmq-discuss] Daemon processes, sporadic messages, and heartbeats?

Tue May 24 23:45:14 BST 2011

Hi:

I'm using python with carrot to connect to RabbitMQ. I'm finding in some cases that long-running processes can "wedge" after they don't get messages for a long time. They don't throw an error and it looks like they are waiting for messages but they never get them. I guess they are connected to the local socket but the server connection is closed (either because the TCP/IP connection times out or the VM sleeps, etc). Once the scripts are killed and restarted they receive messages again.

Is there a heartbeat built into the protocol I can use to keep the connection alive? Is this a bug in the client libs (the fact that they don't throw?).

Has anyone had success running long-running processes (> a couple of days with no messages) with carrot/python (default backend)?

Thanks,
Christian