admin@glassfish.java.net

v3.1 clustering design doc

From: Jennifer Chou <jennifer.chou_at_oracle.com>
Date: Tue, 09 Mar 2010 22:51:31 +0000

Hi Bill,

I was reading over the v3.1 clustering design spec.
http://wiki.glassfish.java.net/Wiki.jsp?page=V3.1Clustering
In general the ideas make sense to me. So the main difference between
v2 and v3.1, is absence of node agent. It will be good if we can note
other differences too.
It sounds like the synchronization mechanism is new in v3.1?

I like the idea of DAS failover.

For v3.1, we expect glassfish product to be installed on each machine
(node) that participates in cluster right?
Then I was confused by the sentence under Data Synchronization - 'In
order to avoid the need to install the software on each node, we would
need to be able to synchronize a)' a) is server software (modules and
libraries)

I'll try to read the v2 cluster arch too to understand better what was
in v2. Is this a good place?
http://wiki.glassfish.java.net/Wiki.jsp?page=GlassFishV2Architecture

Thanks,
Jennifer