Clustered Service MBean with @Clustered, @Service(..)

[db__]

Is it possible to define a clusterwide singleton service, through

@Clustered
@Service(...)

?
I wrote such a mbean but the service is only a singleton on one clusternode. not clusterwide.
If this feature isn't available is it possible to modify this lookup code to ask a remote jboss server?

try {
MBeanServer server = MBeanServerLocator.locate();

count = (Counbean) MBeanProxyExt.create(Countbean.class,
"com.mine:service=countbean", server);
} catch (Exception e) {
e.printStackTrace();
}

regards,
-Dennis

[bill.burke]

currently there is not yet replication behind @Service.

What you can try is to deploy the .ejb3 archive with your @Service within the

server/all/deploy-hasingleton

And, you do your lookup's through HA-JNDI, that should do the trick....

Let me know if that works.

Bill

[db__]

how do i lookup through hajndi,
I actually call MBeanServerlocator... does this use jndi.properties? to set the lookupserver to port 1100?

[bill.burke]

No

Same properties, but theprovider URL would be localhost:1100

[db__]

so I have to use ctx.lokkup() code instead of mbeanserverlocator right?

[db__]

i dont know how to actually get that mbean it's not bound to an jndi name or i don't see it...
maybe i have to add an annotation to bind it to an jndi name actually?

[bill.burke]

@Service beans can have @Remote, @Local and/or @Management interfaces. If you supply a @Remote or @Local interface it will be bound in JNDI under the name of the interface. Look in the tutorial for more information @RemoteBinding

[db__]

yes got that @remote binding this works great. lookup through hajndi works great.
but deploy-hasingleton doesn't work. its somewhat broken here...
i deployed it on one server and can access it through hajndi from all jboss server.

[bill.burke]

You should be able to access the service on all jboss servers...Or, are you saying that it is deployed on more than one server???

[db__]

nono ;-)
I deployed the .ejb3 archive to /deploy on ONE server and my .sar and .war archive on all servers.
So right now I can access my mbean through hajndi on all server. because the .jsps and class files are on all servers and they can lookup through hajndi.
But it would be nice to deploy the .ejb3 to /deploy-hasingleton so it decides itself on which server to deploy the .ejb3 and of cause does the redeploying to an other server it that one actually serving the mbean is down.
But i can't get /deploy-hasingleton work. if i copy my .ejb3 file in this directory then nothing will happen, even if I try this on fresh installed jboss 4.0.2.

[sappenin]

Has the status of this changed?

Does the @Stateless annotation now work with the @Clustered annotation, or should we still be doing it the way it's described in the above post?

Thanks!

David