Hi,
I just set up a cluster with a DAS on server1, and 2 instances on server2 and 3 respectively. They are using glassfish in memory replication to do session failover.
I have a few questions about this. I first setup the availability service with these settings:
Persistence Type: replicated
Persistence Period: web-method
Persistence Scope: session
I setup a loadbalancer using Sun web server 6.1 and the glassfish plugin.
with the configuration above the system was just too slow. I changed the Persistence Scope to both "modified-session" and "modified-attribute", but the were both very slow. I changed the Persistence Period to "time-based" and the performance was alot better.
But, even with this setup, direct connecting to one of the cluster instances are alot lot faster.
Are there any guidelines for doing this sort of setup?
Also I have read about setting up a HADB, is this available in Glassfish?
Are there any documentation on this? And how it actually works?
What is the role of the DAS?
I shut the DAS down and session replication was still working. Is the DAS involved in the session replication at all? Or is that dealt with by the node agents or the cluster instance itself?
Thanks for you help
Hans
[Message sent by forum member 'hansflhotmailcom' (hansflhotmailcom)]
http://forums.java.net/jive/thread.jspa?messageID=228361