Last night after it died again, I changed the usageMemory manager settings to 256 MB. It's been up the whole night. For completeness sake, here's the Java process that's running:
/usr/java/jdk/bin/java -Xmx512M -Dorg.apache.activemq.UseDedicatedTaskRunner=true -Dderby.system.home=/opt/activemq-4.1.0/data -Dderby.storage.fileSyncTransactionLog=true -Dcom.sun.management.jmxremote -Djavax.net.ssl.keyStorePassword=password -Djavax.net.ssl.trustStorePassword=password -Djavax.net.ssl.keyStore=/opt/activemq-4.1.0/conf/broker.ks -Djavax.net.ssl.trustStore=/opt/activemq-4.1.0/conf/broker.ts -classpath -Dactivemq.home=/opt/activemq-4.1.0 -Dactivemq.base=/opt/activemq-4.1.0 -jar /opt/activemq-4.1.0/bin/run.jar you can see it's 512MB for heap, and that's reflected in JMX as well. Current the broker memory in JMX shows max at 256MB as specified, and memory usage is 19%, up from 3% when it started out. I'm still concerned that this is just going to continue climbing...it seems like memory's not being released somewhere! We're not using persistent messages, so I can understand that more memory is being used for journaling. But shouldn't this be getting GCed??? I'd be a real pain to have to monitor the memory use via JMX for the broker, and have to restart the broker every 24hours. Christopher G. Stach II wrote: > > > There's an issue open for the waitForSpace issue. I'm not sure if it > applies to your situation. Anyway, you may want to look into the > usageManager settings in the broker configuration to set the limit > higher. A 20MiB default seems weird. I didn't think that there was a > default. > > -- > Christopher G. Stach II > > > -- View this message in context: http://www.nabble.com/Non-Persistent-Topic-not-getting-messages-tf3264593s2354.html#a9099799 Sent from the ActiveMQ - User mailing list archive at Nabble.com.
