If I am not mistaken, by default, the AMQ Message Store and KahaDB issue asynchronous and synchronous writes to disk, respectively. If you reconfigure KahaDB to issue asynchronous writes, you’ll find that its throughput numbers are closer to that of AMQ Message Store’s. However, keep in mind that asynchronous writes are less reliable than synchronous writes.
See 'enableJournalDiskSyncs' in http://activemq.apache.org/kahadb.html Also refer to the following in activemq-throughput.xml <!-- Use KahaDB for persistence Tune it a bit so we minimize IO operations For more information, see: http://activemq.apache.org/kahadb.html --> <persistenceAdapter> <kahaDB directory="${activemq.base}/data/kahadb" enableJournalDiskSyncs="false" indexWriteBatchSize="10000" indexCacheSize="1000"/> </persistenceAdapter> Joe http://www.ttmsolutions.com/ -- View this message in context: http://activemq.2283324.n4.nabble.com/KahaDB-Sync-Writes-tp3172884p3172994.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
