Re: [infinispan-dev] JMX stuff in infinispan

Galder Zamarreno wrote: > >> <SNIP /> > > Not sure that should be allowed. JMX implementations do moan (via a > InstanceAlreadyExistsException) when the name already exists, why > shouldn't we? You know have the opportunity to give it a different > name (side: we couldn't with JBC - https://jira.jboss.org/jira/browse/JBCACHE-1485) > , so people should make use of it IMO. This is the default behavior: if a second instance is stared under the same domain, then an exception is being thrown indicating that (during cache startup). At this point the user can change domain name, or just enable "forceAutoIncrement" (which is "try using the configured domain, but if it is already taken try using the next one"). Either way, the user *is notified* that a domain conflict exist and he *must take action* in order to avoid that - nothing done under the hood.