This just happended to our testing cluster. It probably deserves a JIRA issue but I am not sure how reproducible it is.
We are using JBoss 4.0.3. SP1 on Windows 2003 Server. Today we updated our 3 node cluster with this hotfix KB913446. For some reason only two machines ( 1 and 3) were updated. Suddenly they started throwing
2006-02-16 11:13:15,802 WARN [org.jgroups.protocols.pbcast.ClientGmsImpl] handleJoin(t01:1139) failed, retrying
Rebooting, restarting, etc didn't stop this. Nor did removing cluster-service.xml. JBoss was NOT accepting any connections so the machines were unusable.
After rebooting machine 2 which did not have the hotfix applied, everything returned to normal. We'll be monitoring this for a while of course.
I hope its just a one off gremlin.
