(Using Jboss 4.0.4 RC2, Windows 2003, JDK 1.5, SQL Server 2000 with Microsoft JDBC Drivers SP3 used for JMS persistence)
We are currently testing JBossMQ clustering of Queues and have been quite successful in gettings things up and running. We have 2 instances of JBoss setup in a cluster (running the all config) with a clustered JMS queue (with quite a few messages in it). An MDB is deployed to each server that drains messages from the queue. All this seems to work well until I restart the database. JBossMQ does not seem to reconnect to the database when the database is restarted. If I restart the DestinationManager (via jmx-console) associated with queue then everything recovers and runs as per normal. If I restart the database again JBossMQ recovers without any user intervention.
I guess my questions are
- can anybody suggest ways in detecting JBossMQ database failures and to allow for automated recovery when the database finally restarts
- why does JBossMQ not recover from the first database restart, but has no problem with the second database restart
BTW, If I do not startup the second JBoss instance, JBossMQ *always* automcatically recovers from a database restart.
So nobody else has seen this problem?