Hi again Bill,

Aha. I was hoping this wasn't going to be the answer. Oh well.

[ snip ]

If this is the case, why does the code for the Valve recommend implementing the Lifecycle interface ? What was the design reason for
that ?h
By implementing Lifecycle, you get well-defined states to allocate and release any resources that your Valve may need. Since the order of
setting attributes is undefined, it makes it much easier to determine
resources that depend on multiple attribute values. It's pretty much
independent of the MBeans.

Excellent, so the Valves can then conform to an defined
servlet/applet/midlet like behaviour.  I noticed the use of the
start() etc methods in the Valve and guessed that was something to
do with it, but also guessed wrong that it was relating to the MBeans.

It's purely because of the Valve definition in server.xml, when it can't
 find the MBean to manage it.  Actually, the Valve should function fine
even with the error:  It's just noise in the log.
Yup.  My Valves were all working just fine with the errors, but I wanted
to be neat and tidy ;-)


The reason I ask is because it seems bit of dicking around just to implement a new Valve. Writing the code took about 2 hours, it then
took 2 days to get the exception stuff sorted out.

What is the additional information used for ? And was there a better (read simpler and less mucking around) way to solve my "get rid of the
exception" than addition of the descriptor, or is that required for
all additional components in Tomcat ?

With the MBean info, it is possible to use other JMX-enabled tools to
manage Tomcat (including your Valve).  The admin web-app is only one
example of such a tool.  It is also likely that JMX support will improve
in Tomcat 5.x.

Of course, the simplest way to "get rid of the exception" is to disable
(aka comment out) the MBeans Listeners in server.xml.
OK.  I'll have a play around with that.  I think I tried that originally,
and when I did, the Administration application stopped working, but that
may just have been co-incidence.

Thanks very much Bill and Craig for your patience.

Cheers,
	-- jon

--
Jon Eaves <[EMAIL PROTECTED]>
http://www.eaves.org/jon/


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to