Since https://issues.jboss.org/browse/ENTMQ-1064 has been introduced customer is noticing delays when using priority messages in the broker with JDBC.For testing purposes I added some logging immediately before and after the executeQuery() call in
org.apache.activemq.store.jdbc.adapter.DefaultJDBCAdapter#doRecoverNextMessages(). Using these just to give an approx. on query execution.| |

Using a local instance of MySQL for jdbc persistence store, with 2 destinations on the broker and 1 million message on each destination. The following timestamps are logged when *prioritizedMessages="false"*| 2015-09-10 17:08:46,559 | INFO | About to call executeQuery() | org.apache.activemq.store.jdbc.adapter.DefaultJDBCAdapter | ActiveMQ BrokerService[localhost] Task-2|

when *prioritizedMessages="true"* the timestamps logged are| 2015-09-10 17:16:58,428 | INFO | About to call executeQuery() | org.apache.activemq.store.jdbc.adapter.DefaultJDBCAdapter | ActiveMQ BrokerService[localhost] Task-21|

With prioritizedMessages="false" seems to take approx 0.2 of a second where the prioritizedMessages="true" is taking about 15 seconds.The prioritizedMessages="true" query execution time does appear to be effected by the number of message in the DB. In my setup with 600,000 messages in the DB it took approx. 5 seconds and for 2 million messages in the DB it took approximately 15 seconds.The environment is only my local machine but IMHO it does demonstrate the difference between prioritizedMessages="false" and prioritizedMessages="true"Can the prioritizedMessages="true" query be speeded up?