users@glassfish.java.net

Re: Glassfish Clustering across 2 Nodes and In Memory Replication

From: John Clingan <John.Clingan_at_Sun.COM>
Date: Wed, 24 Jun 2009 15:02:14 -0700

Correct. This is true for pretty much any application server product.
Something has to front-end the cluster to know to re-direct the
request from a failed node to a running node.

As a side note, we also have customers front-ending a cluster using a
hardware load balancer. So your choice is really mod_jk, the glassfish
load-balancing plugin for IIS/Apache/SunWebServer, or a hardware load
balancer.

John Clingan
GlassFish Group Product Manager

On Jun 24, 2009, at 2:56 PM, glassfish_at_javadesktop.org wrote:

> So even with the fix mentioned by John, in order to support multiple
> physical nodes in a cluster and maintain in memory replication
> across each node, it is a requirement to front the nodes with a Web
> server and the LB plugin.
>
> I'm guessing this is not a requirement for the single node cluster?
> [Message sent by forum member 'radiusmax' (radiusmax)]
>
> http://forums.java.net/jive/thread.jspa?messageID=352810
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>