[rabbitmq-discuss] Clustering failure

[Simon MacMullen]

On 13/12/11 09:24, Dan Wise wrote:
> Can both nodes be disk nodes (I always assumed not)?

Yes, they can. You can have as many disc nodes as you like. However...

 > My question is, is there a way of changing this configuration so that it
 > does not matter which order the nodes come back up after a total outage?

...it can still matter which order the nodes come back. If you shut down 
in an orderly way and A goes down after B, A will need to come back up 
again first (since things maybe have happened in the cluster while B was 
down).

I'm not sure what happens in the event of several nodes simultaneously 
losing power though.

> We cannot rely on
> the fact that the disk node will be usable after such a failure.

Absolutely. You need more than one then :)

Cheers, Simon

-- 
Simon MacMullen
RabbitMQ, VMware