geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jeremy Boynes" <jer...@coredevelopers.net>
Subject RE: [XML][Deployment]POJO design?
Date Wed, 10 Sep 2003 03:35:21 GMT
> From: Sagenschneider, Daniel A 
> [mailto:Daniel.Sagenschneider@team.telstra.com] 
> Sent: Tuesday, September 09, 2003 8:01 PM
> To: geronimo-dev@incubator.apache.org
> Subject: RE: [XML][Deployment]POJO design?
> 
> 
> In regards to configuring EJBs within the Geronimo framework 
> - what consideration is their for configuring (tweaking) the 
> EJBs while the application is running?  

I'm going to <snip/> here as I think this is a different issue.

The POJOs are used to represent the configuration of an object during
deployment (configuration phase). They provide a data model of the
deployment descriptor that is separate from its physical form (e.g. the
precise XML format, whether it is binary, in JNDI, in a RDBMS).

Once the system is running, the runtime attributes can be tweaked
through JMX and a management application. Some things will be tweakable
(e.g. pool sizes, cache state), other things will require redeployment
(e.g. changes to implementation code, assembly information).

The MBean state (its attributes) is persistable (in a form not yet
defined) to allow tweaks to remain after server restart.

Converting the POJO model to MBeans with appropriate attributes is the
job of a different bridge. For example, there is a
ComponentContextBuilder whose job is to convert information in the POJOs
to a JNDI Context instance that can be bound to java:comp for a J2EE
component.

There will be other Builders which configure the components and their
attributes based on the deployment information. This is part of the
Inversion of Control.

--
Jeremy


Mime
View raw message