jmx-dev Bug with jmxmp

[Daniel Fuchs]

Hi Fabrice,

If I remember well this is by design. It the same with all
JMX connector servers. If an application starts a connector
server then it must close it explicitly before exiting.

Sometime this is problematic - especially if you want to
start a connector in a java agent (premain).
A long time ago I wrote an article on that:
https://blogs.oracle.com/jmxetc/entry/more_on_premain_and_jmx

best regards,

-- daniel


On 31/10/14 12:48, Fabrice Bacchella wrote:
> I found a bug with the current implementation of jmxmp. The javax.management.remote.generic.GenericConnectorServer create an anonymous, non-daemon thread.
>
> So it can prevent a jvm to stop.
>
> I found the code at :
> http://grepcode.com/file/repo1.maven.org/maven2/org.jvnet.opendmk/jmxremote_optional/1.0_01-ea/javax/management/remote/generic/GenericConnectorServer.java#GenericConnectorServer
>
> Look at the private class Receiver, it as no custom constructor, and it usage is :
> 289    // start to receive clients
> 290    receiver = new Receiver();
> 291    receiver.start();
> 292}
>
> Is there any bug trafic or official repository for that ?
>
> https://opendmk.java.net is almost dead.
>