Thanks for the question. GlassFish offers two high availability options:
1) In-memory replication: A very fast, robust, easy-to-configure high-
availability option. Here's a screencast that covers it from GlassFish
download to deploying/testing a clustered application in 10 minutes:
http://blogs.sun.com/jclingan/entry/glassfish_clustering_in_under_10
. A vast majority of deployments will use this option.
2) HADB offers 99.999% availability, and covers all the edge cases
(ex: all nodes in a cluster fail and come back). The trade off is that
this option doesn't perform as well as in-memory replication and also
requires additional administration. HADB is not open source, but is
available for download :
http://www.sun.com/software/products/appsrvr/get_it.jsp
(HADB bundle). HADB can be used in production without a subscription
(although if you have 99.999% requirements, we recommend a
subscription. Very few deployments need HADB, but when you need it,
you *really* need it :-)
On Mar 12, 2009, at 12:07 AM, glassfish_at_javadesktop.org wrote:
> Hi All,
>
> I am brand new to the GlassFish and kinda just about downloading.
> I was told some time back that GlassFish used to have clustering
> based on HADB component which is not open source.
> Is this true? Are their any open source alternatives?
> If yes, why shld I use HADB or the other alternative.? basically
> pros and cons.
>
> many thanks in advance.
> [Message sent by forum member 'technologyguy' (technologyguy)]
>
> http://forums.java.net/jive/thread.jspa?messageID=336557
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>