Hi, mk111283, thank you for responding.
I am able to reproduce the bug even with the --asyncreplication=false deploy
option and only accessing the application through instance1, i.e. without
accessing instance2 at all since the cluster was started. In the instance1
log I 'm seeing the TimeoutWarning and in the instance2 log I see the
following warning appear many times:
WARNING|glassfish3.1|org.shoal.ha.cache.command.save|_ThreadID=28;_ThreadName=Thread-1;|SaveCommand
Sending SIMPLE_ACK|#]
I spent a couple of hours trying find some clues when and why this behavious
occurs but in vague. The only thing I can maintain with almost certainty is
that there are periods when the issue is not occuring and everything works
smooth. In my environment those good periods may continue as long as
10minutes. I am able to make about 50 requests without losing the session and
then something goes wrong and session is renewed for almost every second
request. When these "good periods" occur and when they end is a complete
mystery to me, (although even in the "good periods" the warnings in the logs
still appear.)
--
[Message sent by forum member 'spg82']
View Post: http://forums.java.net/node/780204