Aha, thanks for the tip Matthew.<br><br><div name="sig_185a028ade" style="margin-top: 0pt; margin-right: 0pt; margin-bottom: 0pt; margin-left: 0pt; "></div><div class="gmail_quote">On Wed, Jun 16, 2010 at 12:19 PM, Matthew Sackman <span dir="ltr"><<a href="mailto:matthew@rabbitmq.com">matthew@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 Wed, Jun 16, 2010 at 12:02:08PM -0400, Ilya Grigorik wrote:<br>
> One of our rabbit servers is ~7GB RAM, and I know that if we push over 250K<br>
> messages into it (heavy messages), and the server hits swap, the server<br>
> starts crawling.. to the tune of being able to get only a few messages a<br>
> second (out). Meaning, we need to nuke the entire queue to bring the server<br>
> back to functional.<br>
<br>
</div>You need to be using the new persister branch - compile from source, but<br>
set the server to use branch "bug21673". The raison d'être of the new<br>
persister is to solve problems such as the one you describe.<br>
<font color="#888888"><br>
Matthew<br>
</font></blockquote></div><br>