5.1.0. I believe it's happening with 5.2.0, too, but have not thoroughly tested.
The problem is that the broker runs out of memory. At the point in time of the failure, no more messages could be sent anyway because the broker is out of memory. So we're less worried about blocking producers and more worried about how to maintain the health of the broker. We set the message queue limits to prevent more than X messages on a topic for each consumer. The queue limits are only set to prevent the broker from running out of memory if a particular consumer starts blocking. We are hoping that if a consumer dies, the queue limits will at least keep any message queues for the dead consumer from going out of control. -----Original Message----- From: Bruce Snyder [mailto:[EMAIL PROTECTED] Sent: Monday, December 08, 2008 10:17 AM To: users@activemq.apache.org Subject: Re: Broker not releasing memory On Mon, Dec 8, 2008 at 7:28 AM, Frugia, Kirby A <[EMAIL PROTECTED]> wrote: > I can reproduce this with the non-embedded, out-of-the-box broker. I also > think it's an under-the-hood problem. > > Until we get a fix, we've set pending message queue limits on our busy > topics. So, for example, the broker will only queue up 500 messages on a > particular topic per consumer. This will prevent things from getting too out > of hand on a single client crash, but is not the solution to the problem. What version of ActiveMQ are you using? Also, have you disabled producer flow control to allow the producer to continue to send even when the consumer is going very slow? Bruce -- perl -e 'print unpack("u30","D0G)[EMAIL PROTECTED]&5R\"F)R=6-E+G-N>61E<D\!G;6%I;\"YC;VT*" );' Apache ActiveMQ - http://activemq.org/ Apache Camel - http://activemq.org/camel/ Apache ServiceMix - http://servicemix.org/ Blog: http://bruceblog.org/
