[rabbitmq-discuss] Installing broker on each node instead of client-only approach

Jason McIntosh mcintoshj at gmail.com
Tue Apr 15 16:06:02 BST 2014


Just a note - we use local shovels on the distributed nodes instead of
federation on the central node so that the "central" rabbitmq server
doesn't require knowledge of the remote nodes, e.g. cloud computers which
can come up and down as you wish, and the central piece doesn't have to
know about them, it just sits there waiting for data.

Jason


On Mon, Apr 14, 2014 at 9:30 AM, Matthias Radestock
<matthias at rabbitmq.com>wrote:

> On 14/04/14 15:04, Николай Максимов wrote:
>
>> Thank you, Matthias. Is there any limits for queue size to be reached
>> while working in autonomy/degrade mode? I need this to decide how
>> long my setup could work without connectivity.
>>
>
> The limiting factors are memory and disk space. When memory gets tight and
> messages are fully paged out to disk, each message requires about 122 bytes
> of memory (on a 64-bit machine), and about 255 bytes + payload size of disk
> space.
>
>
> Matthias.
> _______________________________________________
> rabbitmq-discuss mailing list
> rabbitmq-discuss at lists.rabbitmq.com
> https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss
>



-- 
Jason McIntosh
https://github.com/jasonmcintosh/
573-424-7612
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20140415/303e9ce0/attachment.html>


More information about the rabbitmq-discuss mailing list