<div dir="ltr"><div class="gmail_extra">On Thu, May 8, 2014 at 3:46 AM, Simon MacMullen <span dir="ltr"><<a href="mailto:simon@rabbitmq.com" target="_blank">simon@rabbitmq.com</a>></span> wrote:<br></div><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"><div class="">On 08/05/2014 05:26, Garrett Johnson wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hello all — I have a question about making sure a message NEVER hits<br>
disk. Assuming the messages are not published as persistent (duh) and we<br>
cancel connections at a given memory threshold rather than paging to<br>
disk, what else is there to keep in mind?<br>
</blockquote>
<br></div>
Assuming this is broker-wide, you would want to set vm_memory_high_watermark_<u></u>paging_ratio to some large number. You should not need to cancel connections if you do that. </blockquote><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
</blockquote><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class=""><br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
If it is even possible, can you still mirror the queues in a cluster?<br>
</blockquote>
<br></div>
HA is still fine; it's orthogonal to persistence.<br></blockquote><div><br></div><div><br></div><div>Okay cool, even when a new node joins and has to sync?</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
Cheers, Simon<br>
<br>
</blockquote></div><br><br clear="all"><div><br></div><br><div dir="ltr"><span><div><br></div></span></div>
</div></div>