[rabbitmq-discuss] vhostpath

Steven Taylor taylste at gmail.com
Mon Mar 21 11:52:16 GMT 2011


if an exchange / queue is defined in a vhost (being a distinct namespace),
are there other benefits such as:

when a new queue is created, it wont cause a reparse / graph node rebuild of
*all* routing keys, only those in this particular vhost/namespace?  i.e.
RabbitMQ (can't remember exact Rabbit terminology) use a topology/graph
based search algorithm to make topic exchanges message destination
resolution more efficient.

Is this statement True or False?  Not sure how important this is.  Not sure
how expensive the routing graph rebuild thing is... it just occured to me
while thinking about a certain senario.

Also, thanks for the info... I know you guys live and breathe the AQMP spec,
but most of us aren't going to read it if we don't have to.  I think many
newcommers will have this perspective.  That said, I do plan on looking at
the AQMP spec at some point.


On 21 March 2011 09:53, Emile Joubert <emile at rabbitmq.com> wrote:

> Hi Steven,
> On 21/03/11 08:02, Steven Taylor wrote:
>> Hi,
>> I can't find a definition on rabbitmq.com <http://rabbitmq.com> for
>> vhostpath
> [...]
> Thanks for pointing this out. The AMQP specification has details in section
> 3.1.2:
> """
> A virtual host comprises its own name space, a set of exchanges, message
> queues, and all associated objects. Each connection MUST BE associated with
> a single virtual host
> """
> At the time of writing the specification was here:
> http://www.amqp.org/confluence/download/attachments/720900/amqp0-9-1.pdf
> Regards
> Emile
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20110321/7db24120/attachment.htm>

More information about the rabbitmq-discuss mailing list