[ataylor]

Since the backup servers must be started before any live server - for  whatever reason - there seemed to be a hen-egg-problem with this  configuration

This will no longer be an issue with the next version.

It would be great if any JBoss "all" configuration had a preconfigured  live and backup server and the backup server would use autodetection by  default to find their live servers. And of course, if a failed live  server comes back, no prequisite restarts of other server should be  needed

Difficult, when shared store is used the backup server will need the location of the journal configured and if Replication is used then other config will also need to be added. However, we will be providing a template with extensive docs.

[ataylor]

Sebastian, I definitely agree.  I would like to be able to define both a  live node and a backup node (for another server) in a single  hornetq-configuration.xml.  Defining the backup node would be optional.   If defined, 2 instances of HornetQ would be started (under the hood).   The fact that two instances were started would be somewhat transparent  to someone deploying a HornetQ implementation.

Adding another server to a hornetq-configuration.xml file does not really make sense, this file represents a single server and its configuration, each server will have its own journal, acceptors, connectors and bridges etc. A server is actually represented by the hornetq-jboss-beans.xml file, take a look, and each is configured with its own hornetq-configuration.xml. To add a backup server, or another live, you would just create a hornetq-jboss-beans.xml which references its own configuration.

This proposal assumes there is nothing in the internal design of HornetQ  that prevents two instances running in the same JVM (so long as  separate ports and journal-directories can be specified in the  configuration file.

Currently the only thing that prvents this, within the AS that is, is how jms resources are added to jndi and how objects are added to JMX and to the micrcontainer internally. At present there will be a naming clash, however again we are working to deal with this issue for the next release.

Basically, each live server will have the JMS resources configured, each back up will just have a basic server configuration, the JMS resources will be recreated on the backup server on failover via the shared store or replication (when added). The backup server will have a bridge defined to live using an invm connector for each destination configured. In later releases we will probably automate all this, currently time and resources dictate what we can get done.

[sebastian.koske]

Hi,

I have just come across the new version. As of now it seems that really all the problems I had were solved:

http://docs.jboss.org/hornetq/2.2.2.Final/user-manual/en/html/clusters.html

Plus some extra cool stuff like Failing Back To Live Server. Thanks a lot for the great work, looking forward for the AS 7 integration.