Hi,
We are trying to integrate Shoal since a few weeks into our web
applications deployed on a Glassfish 2ur1 cluster (2 instances so far).
Our main purpose is to be able to determine which instance of the
cluster is the current "master" in order to perform a one time task.
While using the version 1.0-11142007, the Shoal cluster was unable to
elect a new master after an application redeploy OR undeploy/deploy
almost each time. With the version 1.1-03242008 using jxta 2.5, we are
still experiencing the same problem but less frequently (once each
10/15 times).
We tried to work around this problem with a basic admin interface
allowing us, for each instance, to shut it down or join a group.
However, after a shutdown/join (with the famous
"MASTER_CHANGE_EVENT, ..." log), a call to
"getGms().getGroupHandle().isGroupLeader()" returns false on each
instance.
Did you already run into this problem before ? Could it be linked to
the usage of Shoal by the Glassfish cluster ?
Thanks in anticipation,
Regards