How to implement get/setDistributedState?

[timdrisdelle]

I've got a singleton mbean (extends HASingletonSupport), and I want to have some runtime-configurable parameters be shared across the cluster.

If a parameter is changed on one of the nodes, I want it to be reflected across all of the nodes (...which is why I'm using singleton!).

I read the Clustering doc, but I still can't figure out how I'm supposed to implement the getDistributedState and setDistributedState methods in my mbean.

should I still have a private member variable in the mbean, with a getter and setter?

or should there be no private member variable, and the getter/setter call the get/setDistributedState methods?

And how should the get/setDistributedState methods work with the start/stopSingleton methods?

Vexed in Victoria,
Tim.

[slaboure]

I would instead use the DistributedState service: no need to implement anything, just use it as is, very convenient to share configuration accross a cluster.

cheers,

sacha

[timdrisdelle]

My mbean extends HASingletonSupport, which extends HAServiceMBeanSupport.

The HAServiceMBeanSupport class has two "convenience" methods called getDistributedState and setDistributedState, which both delegate to the DistributedStateService.

Therefore, I should be able to call getDistributedState and setDistributedState from within my mbean.

My question is really, how and when do I make these calls? Do I still need individual getters and setters for my runtime configuration properties, each of which will call get/setDistributedState?

Tim