Some usages of the DefaultThreadFactory in the clustering subsystem lead to thread groups being created repeatedly on server reload.

A thread group is created each time a DefaultThreadFactory instance is created, so it depends on how the DefaultThreadFactory instances are handled. If they are kept in static fields, both the DefaultThreadFactory and the ThreadGroup instance is only created once. If the DefaultThreadFactory instances are created as part of operation that can be executed multiple times in the server lifecycle, that leads to ThreadGroups accumulating.

Problematic places are these:

• CacheRegistry#topologyChangeExecutor field,
• CacheContainerServiceConfigurator#get() method,
• InfinispanSessionManagerFactory#executor field.