]
Daniele Pirola commented on WFLY-5189:
--------------------------------------
I found a solution: inside {{discovery-group}} and {{broadcast-group}} instead of specify
{{jgroups channel}} I specify {{socket-binding}}, binded to default {{jgroups-mping}}
binding. Inside {{tcp}} stack I use {{TCPPING}} instead of {{MPING}} protocol so this
binding wasn't in use anymore inside my configuration.
Eliminate "discarding discovery request for cluster=X from Y;
our cluster name is Z" WARN messages
--------------------------------------------------------------------------------------------------
Key: WFLY-5189
URL:
https://issues.jboss.org/browse/WFLY-5189
Project: WildFly
Issue Type: Bug
Components: Clustering, JMS
Affects Versions: 10.0.0.Beta2
Reporter: Paul Ferraro
Assignee: Paul Ferraro
Fix For: 10.0.0.CR3
These WARN messages are due to the fact that WF no longer uses a shared transport for its
jgroups channels (JGroups recently deprecated shared transport in favor of fork channels).
While these messages are harmless - they are quite annoying - and are the subject of
several forum posts.
Currently, the messaging-activemq subsystem uses JChannels configured from a specific
protocol stack. We should modify the default configuration to use a ChannelFactory from
the default channel, which will use a new fork of the default "ee" channel. In
xml/dmr terms, this corresponds not providing a value for jgroups-stack.
N.B. At the moment, this change will introduce new annoying WARN log messages until
JGroups is upgraded to 3.6.5.Final, hence the jira dependency.