[jboss-as7-dev] DOCS: Researching details for standalone clustering

[Jason Greene]

Correct. Right now the only thing you can do is use infinispan and jgroups directly. However the ininispan in beta3 has issues running in a modular env (there is a patch for this in 5cr1). The most recent checkout works around the problem.

Sent from my iPhone

On Apr 28, 2011, at 10:09 AM, Andrig Miller <anmiller at redhat.com> wrote:

> 
> 
> ----- Original Message -----
>> From: "Darrin Mison" <dmison at redhat.com>
>> To: jboss-as7-dev at lists.jboss.org
>> Sent: Wednesday, April 27, 2011 7:24:58 PM
>> Subject: [jboss-as7-dev] DOCS: Researching details for standalone clustering
>> Hi, I'm chasing down details about the "standalone clustering" (as
>> included in Beta3) for documentation.
>> 
>> I've noted the following:
>> 1 - There's a clustering-standalone.xml configuration file in
>> standalone/configuration that you can start the server with using
>> ./standalone.sh -server-config=clustering-standalone.xml
>> 
> 
> Actually, no.  It turns out that the command-line argument handling in Beta 3, requires this syntax:
> 
> ./standalone.sh -server-config clustering-standalone.xml
> 
> The "equal" sign does not work, but a space does.  I believe Jason hinted that he might fix this, so it may change in Beta 4.
> 
>> 2 - This configuration differs from the standard config in that:
>> - the org.jboss.as.clustering extension is enabled
>> - some additional items are configured in the
>> urn:jboss:domain:osgi:1.0 subsystem
>> - 2 new subsystems configured (jgroups and infinispan)
>> - 6 new socket bindings added to the standard-sockets group for
>> jgroups
>> 
>> So exactly what feature set does this give you out of the box for
>> clustering?
> 
> So far, nothing that I can tell.  I just had a e-mail conversation with Remy on what I thought might be an issue with HTTP session replication, but he says its not in there yet.
> 
> Also, while there is a configuration for Hibernate 2LC, that also is not completely implemented.
> 
> My guess is that none of what's configured is actually complete.
> 
>> 
>> I'm guessing session replication between standalone instances with
>> JGroups for intra-node communication and Infinispan for storage?
>> 
>> Anything else?
>> 
>> Other questions:
>> 
>> How does each instance determine which other instances it should be
>> clustered with?
>> 
>> Does this work with multiple instances on one machine (multi-homed or
>> on different ports)?
>> 
>> Can multiple standalone instances using "cluster" configuration live
>> on the same network but not all be one cluster? Like 10 instances as 2
>> clusters of 5 for example?
>> 
>> Aside from centralized management, what are the other key features
>> from domain mode that you won't get from standalone clustering ?
>> 
>> Are there any known changes that are going to be made to this feature
>> set for GA ?
>> 
>> 
>> Thanks in advance.
>> 
>> 
>> 
>> 
>> --
>> Darrin Mison
>> "I'm not a JBoss Developer, I just play one on TV"
>> Content Author - Engineering Content Services
>> Red Hat, Inc.
>> 
>> 
>> _______________________________________________
>> jboss-as7-dev mailing list
>> jboss-as7-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/jboss-as7-dev
> _______________________________________________
> jboss-as7-dev mailing list
> jboss-as7-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/jboss-as7-dev