<div dir="ltr">On Fri, Mar 1, 2013 at 10:12 AM, James Carr <span dir="ltr"><<a href="mailto:james.r.carr@gmail.com" target="_blank">james.r.carr@gmail.com</a>></span> wrote:<br><div class="gmail_extra"><div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hey All,<br>
<br>
We've been re-architecting our infrastructure to be evenly distributed<br>
across availability zones and I'm curious if there are some general<br>
guidelines for clustering rabbitmq across availability zones. I've<br>
been thinking about this quite a bit and am wondering if you can<br>
cluster across the zones (iirc this wouldn't be a good idea due to<br>
network partitioning). Federation comes to mind but then each node in<br>
the federated cluster would get a copy of a message when only one<br>
consumer needs to grab it.<br>
<br>
Anyone have more experience with this?<br></blockquote><div style><br></div><div style>We cluster across availability zones and haven't run into any problems (yet). I think federation is overkill. You'd probably want to use federation if you wanted to cluster across regions. My only recommendation is to run inside a VPC if you can that way you have control over private IP addresses and hostnames. That makes things a lot easier.</div>
<div style><br></div><div style>Bryan </div></div></div></div>