activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "James Strachan" <>
Subject Re: Slave waiting for lock during JBoss Deployment
Date Tue, 13 Mar 2007 09:39:37 GMT
On 3/12/07, tcort <> wrote:
> About my setup:
> I'm running ActiveMQ (4.2 Snapshot) on JBoss (4.0.4) with Java (1.5) on
> Debian GNU/Linux (stable/sarge). I have 2 JBoss instances and 2 MySQL
> (5.0.32) instances. MySQL's database replication is setup and working. I
> have ActiveMQ setup in a JDBC Master Slave configuration.
> What I want:
> I'd like one broker that either JBoss instance can connect to. If the Master
> broker goes down, I want the salve to take over as the Master. If possible,
> I'd like to create one EAR file that can be deployed on both servers.
> What is happening:
> It sort of works. I start one JBoss instance and it is happy. I try to start
> the second instance and the second instance hangs mid-deployment. The last
> message in server.log before it hangs is: "2007-03-12 13:34:57,194 INFO
> [] Database driver
> recognized: [mysql-ab_jdbc_driver]". When I shutdown the first instance, the
> second instance finishes deploying and is happy.
> My Ideas:
> From the log messages, archives I've searched, and documentation I've read,
> it looks like the second instance is waiting for the first instance to
> release a lock.
> The Question:
> Is there a way to have the second instance (the slave broker) create a
> separate thread that waits for the master to fail so that it doesn't hold up
> the deployment of the application? Is there some sort of NOWAIT option? Or
> do you have any suggestions for a different strategy?

Is there a JBoss way to asynchronously start the broker? Failing that
we welcome patches if you want to try change the broker startup code
to make it optionally asynchronous.


View raw message