You are making 2 bold assumptions. 1. JBoss will always run with the EJB service installed. 2. JBoss will always have a database available.
Neither of these hold. As a quick example, I may want JBossMQ with out a database or EJB services. -dain -----Original Message----- From: David Budworth [mailto:[EMAIL PROTECTED]] Sent: Saturday, December 01, 2001 6:04 PM To: [EMAIL PROTECTED] Subject: [JBoss-dev] RequiredModelMBean.java? / general rantings Marc / everyone, When you asked about this Dynamic mbean thing I'm working on, were you thinking of me applying it to RequiredModelMBean? If I read correctly, we are required to supply an implementation of that class, no? If not, ignore the rest. I'd be happy convert my stuff over to be the implementation of that, since mainly the only difference with what I wrote and this is the persistent storage stuff. I do have a few questions on how it should be done. 1) What are the expectations for determining the MBeanInfo? Should we expect a XYZMBean interface to match the XYZ implementation the user provides? (similar to regular MBeans) This would be easy to add. Since I already have the code that walks the methods of any specified interface to compose the operation/attribute info structures. 2) What should be the rules for determining the operations/attributes? I have written and re-written this logic in my own code about 15 times, never really happy with it. Example, how to handle: int getXYZ(); void setXYZ(float); Do you consider the get to be a RO attribute and one to be an operation? Or throw an exception for non-compliant naming? I see nothing in the spec regarding naming standards on dynamic mbean oper/attr or int getXYZ(); void setXYZ(int); void setXYZ(float); Do we consider get/set(int) to be a RW attr, and set(float) to be an operation? Or throw again? In my stuff, there is a protected Class[] getInterfaces() so an mbean can specify which interfaces it want's to expose to management. But, if that returns null(default impl), I just use all public methods not defined in the base or java.lang.Object to be a managed UI. Would you want that in the JBoss RequiredModelMBean? What I have now, basically allows the subclasses to specify any part of the MBeanInfo (ie, via "protected MBeanAttributeInfo[] getAttributeInfo()"), so the subclass can 'break the rules' that are defined in the base class. I wonder, is this too much complexity to offer in a generic base class to be supplied with JBoss? As for persistence, have you finished rolling on the floor laughing at my idea of using EJBs to store? I have noticed that no other components use EJBs for their JDBC based persistence. Is there a reason for this? <soapbox> Assuming Dain's engine has nothing to persist (MBean wise), there is no reason not to use it. If we don't believe in EJBs enough to use them ourselves, how can we tell others to use JBoss for their projects. Hell, configuration persistence is something that happens so rarely (in the application sense), I don't think performance is really an issue. </soapbox> As for clustering stuff, keeping in mind I haven't looked at it yet. Does anyone know if hypersonic get's clustered as well? I see that to get the EJBs mass deployed there is the farm directory, so that kind of implies that there is a master server somewhere (unless everyones farm propogates to everyone else). If hypersonic doesn't cluster, then is there the ability to add to a DD something like: if (master) use local DB else use master DB We store the JMS stuff (and maybe others, I haven't looked) on disk now. With the option of using JDBC. I understand that there is a performance issue, but from the admin point of view, they'd be much happier if everything was all in one place, that was remotely viewable (ie. whatever DefaultDS points to). I've heard Marc mention several times that the winner in this space is going to be the ones with the best ease of management. And I fully agree. Knowing what it took to manage eloan.com's website, and all the problems with what files are where, all the little cron jobs on different machines to make sure the old stuff gets nuked to avoid running out of disk space. I think that a system, with a single point of management (ie: everything in single DB) would have made our lives easier. Keep in mind, my view may be tainted. Our admins pretty much all came from Oracle (or oracle based companies). So to them, having a DB be the front for everything was nirvana. My current project has all my custom config info in the database. So I can easily make changes that all servers can see, as well as perform atomic changes (update 20 config items, then commit, rather than calling setXXX, setYYY and having each change propgate one at a time). But hey, maybe that's just me. Is my view of how things should work just completely misguided? I'll shut up (again) now. -David _______________________________________________ Jboss-development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development _______________________________________________ Jboss-development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development
