removal of slave nodes and network topology

["Dondorp, Erwin" <er...@cgi.com.INVALID>]

Hello users!

I'm using a (test) Artemis cluster with auto-discovery using jgroups+file_ping.
When master- or slave-nodes are added, Artemis shows them immediately after a broker-diagram refresh.
When master-nodes are removed, and after the reconnect-attempts are exhausted, the server forgets the deleted master-nodes and the diagram is also properly updated.
But when slave-nodes are removed, and after the reconnect-attempts are exhausted, the server keeps reporting the deleted slave-nodes in the broker-diagram forever.
This goes on even after restart of the corresponding master nodes.

Is this the expected behavior?
Is there maybe some configuration involved that I missed on when to forget a slave-node?

thx!
Erwin