[rabbitmq-discuss] RabbitMQ Shovel problem

[Gabriel Hernandez]

Hi,

I have a system where I need to not lose any message in any way while processing, so I have a local rabbitmq server where the publisher is located. The messages in these servers are then pulled from a central location where they are consumed.

Producer - >  B Node 1     ---------------- >  A Node (Shovel) --------- > Consumer

For this I have a fairly basic shovel configured in the A Node to pull from the processing servers. At first everything works fine, but after some time of the connection being idle the shovel seems to stop working. When a new message is published to the B Node at first it appears to be sent, but not acked, so in the administration plugin the queue shows: Ready messages: 0  Unacked messages: 1  Total: 1. After a while the message gets requeued and never taken by the shovel. In the A Node messages can be published to queues and exchanges and messages can be consumed, so it's not a problem with rabbitmq itself.

When I try restarting the server using rabbitmqctl it seems to stop the node, but then it hangs and any effort to start or stop hangs, I have to kill the erlang beam process.

I'm using RabbitMQ 2.8.1 with Erlang R14B04 on CentOS 6. These are the latest versions I found RPM's for.

Anyone has any idea what's happening here and what I might be able to do to fix this?




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20120412/6495c4b0/attachment.htm>