[rabbitmq-discuss] Roadmap Questions

Thu Jun 4 14:39:33 BST 2009

Tsuraan,

tsuraan wrote:
>> There are actually some fairly easy ways to get the kind of reliability
>> you refer to from the existing RabbitMQ broker. Just to mention two of
>> the options: 1) get the persister to write to a SAN, such that in the
>> event of a node failure a standby node can pick up from where the first
>> node left off
> 
> is this a trivial thing to do?  I know the queue and binding
> definitions are in a mnesia database, which is hostname dependent.  If
> you had whatever machine that's actively running rabbit always have
> the dns entry "rabbit.example.com", and you always start your erlang
> with -name rabbit at rabbit.example.com, will mnesia always come up and
> have everything working?

Generally yes. Mnesia recovers automatically from most faults, and the
steps for manual intervention in cases where it can't are well
documented. The same goes for the persister, though it may need a little
nudge (renaming the log file).

It's worth noting that the above is designed to cope with the
*occassional* fault - if you have machines falling over every day, or
cannot afford a few minutes outage, then option 2 - the dual queue
approach - is preferable.

Matthias.