<br><br><div class="gmail_quote">On 31 October 2011 13:26, Simon MacMullen <span dir="ltr"><<a href="mailto:simon@rabbitmq.com">simon@rabbitmq.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im">On 31/10/11 13:20, Pete Kelly wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
We are looking at clustering our RabbitMQ setup at the moment - however<br>
we are looking to do it across 2 data centres, one in Europe and one in<br>
the USA. There is a VPN between the 2 DC's so packets can be routed<br>
between the 2 networks directly over the VPN.<br>
<br>
Is this a good idea, or would it be classed as one of the "WAN"<br>
scenarios which is warned against in the Clustering guide?<br>
</blockquote>
<br></div>
No and yes, unfortunately.<div class="im"><br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
If it's is a good idea, great! However if not, we will look into having<br>
clusters at each end with mod shovel or take another approach.<br>
</blockquote>
<br></div>
Another approach would be the federation plugin.</blockquote><div><br></div><div>Thanks, it mentions federation is incompatible with clusters at release 2.5 (I assume), is this still the case?</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im"><br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Are there any further docs or explanations anyone can point me to to<br>
describe the specific problems RabbitMQ has over a "WAN" ?<br>
</blockquote>
<br></div>
It's discussed a bit in this blog post:<br>
<br>
<a href="http://www.rabbitmq.com/blog/2011/06/22/federation-plugin-preview-release/" target="_blank">http://www.rabbitmq.com/blog/<u></u>2011/06/22/federation-plugin-<u></u>preview-release/</a><br>
<br>
but the short version is that its the CAP theorem. RabbitMQ clustering chooses Consistency and Availability and thus loses Partition Tolerance - which doesn't work out so well if the links between nodes can go down.<br>
<br>
Cheers, Simon<br><font color="#888888">
<br>
-- <br>
Simon MacMullen<br>
RabbitMQ, VMware<br>
______________________________<u></u>_________________<br>
rabbitmq-discuss mailing list<br>
<a href="mailto:rabbitmq-discuss@lists.rabbitmq.com" target="_blank">rabbitmq-discuss@lists.<u></u>rabbitmq.com</a><br>
<a href="https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss" target="_blank">https://lists.rabbitmq.com/<u></u>cgi-bin/mailman/listinfo/<u></u>rabbitmq-discuss</a><br>
</font></blockquote></div><br>