RabbitMQ cluster questions



Hi. i need to make rabbitmq HA cluster, with good throughput, and maybe with loadbalancing haproxy. Main task is to prevent data loss. What i made?
I install two nodes on separate machines, make cluster with mirrored queues. So messages from first node sync on second node. And when node down, messages available on the second node. So the questions.
1) Is mirror queues enough for save data when some node will fall down? Or what can i do, for saving messages?
2) If mirror queues is on, so information on hard drive of first node must be replicate on the second node automaticaly? Or no? My second node database folder is empty, but if master falls, messages are still on the second node, and i dont know where are they store.
3)Is there any client for viewing Rabbitmq database? Like aqua? I found an erlang observer, but dont know how to use it.
4) Load balancing i'm planning to make with haproxy on frontend, for example if i push 8000 messages, they must distribute on two nodes? Like 4000 on 1st node and other on the second? Because when i'm sending messages on my nodes i see 8000 messages on each node.
Hope you'll help me. Thank you.

Members online