[rabbitmq-discuss] Queue replication (I know, I know)

[Derek Chen-Becker]

On 10/15/2010 04:48 AM, Simon MacMullen wrote:
>> We don't need HA or real-time failover, we just need to make sure that
>> persistent messages get stored on the broker and backed up somewhere
>> else so that if the broker server goes belly up we can somehow restore
>> what was in the queue at the time that the server failed.
> 
> Ah, in that case you don't even need a backup broker (and hence shared 
> storage). Just redundant storage. Is that practical?
> 

Basically, yes. Ideally, the broker would duplicate every file write
operation to a second location. It would be really nice if we could do
this with Mnesia, too, so that essentially we're getting a real-time
copy of state on a backup machine. If the primary machine went down I
would do the work of starting a new broker on the backup machine using
the redundant files. The only concern I have with this scenario is that
it sounds like RabbitMQ uses the node name to identify itself, so I
wasn't sure if that was configurable or if I could uses the hosts file
to fake it.

Like I said, if we weren't constrained to run on Windows I would just be
doing DRBD and this would all be fine, but the decision on OS was made a
long, long time ago and it's infeasible to change it now :(

Thanks,

Derek