ISPN-5023 Random Failues on *FactoryBeanTest in the Spring Modules

[gustavocoding]

https://issues.jboss.org/browse/ISPN-5023

Wasn't able to reproduce it locally, but I think sometimes there's a clash inside Jgroups Jmx registration, given that the modules run by default on 15 threads. Changed it to 1 Thread, test time is roughly the same in my environment

[anistor]

Looking..

[danberindei]

I don't think we should be calling JGroups' JmxConfigurator.registerChannel(), at least not with a global server.

[anistor]

Integrated. Thanks @gustavonalle !

[gustavocoding]

I don't think we should be calling JGroups' JmxConfigurator.registerChannel()

@danberindei Why shouldn't it be called with a global server? Looking at the javadoc for this method: Registers an already created channel with the given MBeanServer.

[danberindei]

@gustavonalle I thought TestCacheManagerFactory took care of creating a separate MBean server for each test thread, but I see I was wrong and it only sets a unique domain name.

Looks like the ISPN-3125 fix replaced the call to AbstractEmbeddedCacheManagerFactory.createCacheManager(InputStream is) with a direct new DefaultCacheManager, and the domain isn't renamed. I've created #3108 to fix that.