[rabbitmq-discuss] Question on internal guid or hashing msg content

[Jon Brisbin]

On Jan 22, 2011, at 4:52 PM, Matthew Sackman wrote:

> Hi Jon,
> 
> On Sat, Jan 22, 2011 at 04:45:28PM -0600, Jon Brisbin wrote:
>> Can I use the guid property of #basic_message{} for this
> 
> Yes, that's what we do. Check the type sigs of msg_store - the key there
> is always the guid.
> 
>> or would I still get unnecessary duplication (I guess I'm wondering if the same message content can have different guids for sending to multiple queues)?
> 
> The guid is created and applied in the channel on basic.publish. Then it
> is routed to various queues. Thus every queue receiving said msg sees it
> in the same #basic_message record with the same guid.

Another question: since non-clustered nodes will have no visibility of guids from other nodes, is it necessary to store messages separately (i.e. one bucket per cluster or node)? Maybe storing all messages in a single bucket would be easier for out-of-band processing?


Thanks!

Jon Brisbin

        Web: http://jbrisbin.com/
    Twitter: @j_brisbin
      Skype: jon.brisbin

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20110123/6e8878f3/attachment.htm>