<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
<font face="Calibri">Matthew - do you have any idea of when the new
persister will go 'mainstream?'</font><br>
<br>
On 5/19/2010 9:55 AM, Matthew Sackman wrote:
<blockquote cite="mid:20100519145513.GA1896@mrnibble.lshift.net"
type="cite">
<pre wrap="">On Wed, May 19, 2010 at 10:45:19AM -0400, <a class="moz-txt-link-abbreviated" href="mailto:tyler@echonest.com">tyler@echonest.com</a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">The server process does get busy, up to about 80% of the available cpu.
It uses all of the available memory, and crashes shortly after. (Shouldn't
it be pinning the queues to disk automatically, or do i need to use 21444
for that...?)
</pre>
</blockquote>
<pre wrap="">
The new persister will gently push messages to disk when memory gets
tight. That is not the behaviour of default branch or any released
version of Rabbit.
</pre>
<blockquote type="cite">
<pre wrap="">The server never recovers, even if I kill all producers and most
consumers, although the beam.smp process hangs on to all
the memory. I'm playing with 21444 now to see if i can replicate this.
</pre>
</blockquote>
<pre wrap="">
21444 is ancient and should not be used. If you want to use the new
persister, the correct branch is bug21673.
Matthew
_______________________________________________
rabbitmq-discuss mailing list
<a class="moz-txt-link-abbreviated" href="mailto:rabbitmq-discuss@lists.rabbitmq.com">rabbitmq-discuss@lists.rabbitmq.com</a>
<a class="moz-txt-link-freetext" href="http://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss">http://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss</a>
</pre>
</blockquote>
</body>
</html>