[
https://issues.jboss.org/browse/WFLY-6728?page=com.atlassian.jira.plugin....
]
Daniel Fröhlich commented on WFLY-6728:
---------------------------------------
No, it is not the same thing as with EAP6. With EAP6, the use AMQ6 embedded, not via RA.
So the AMQ broker is running inside the same jvm as EAP does. They do so because they want
a very high likelihood that the PUT of a message does succeed. And if Sender and Broker a
within the same JVM, that is the case. In the RAR approach with an external running Broker
(maybe on the same host, but different jvm process), there is a chance that the broker
might be offline (not yet started, killed by kernel oom killer, whatever). That approach
follows good HA principles to reduce the number of components in your design. So again, a
perfectly good architectural decision.
And again that makes it important that we support JDBC persistence not only in AMQ via
RAR, but also embedded in EAP.
They have a standing rule in place to never ever go into production with "x.0.0"
numbers ;-)
They will start the migration from EAP5/7 to EAP7 next year with go live target in 2018.
So there should be enough time for us to implement and stabilize the functionality.
JDBC persistence-store for Artemis
----------------------------------
Key: WFLY-6728
URL:
https://issues.jboss.org/browse/WFLY-6728
Project: WildFly
Issue Type: Enhancement
Components: JMS
Affects Versions: 10.0.0.Final
Reporter: Jochen Cordes
Assignee: Jeff Mesnil
Apache ActiveMQ had the capability to store messages into a database via JDBC. In Apache
ActiveMQ Artemis this has gone.
For a consistent backup data of various (co-located) systems participating in
transactions should reside at the same datastore as otherwise this needs to be achieved
through application software design (i.e. idempotent consumers etc.).
As in Apache ActiveMQ Artemis a JDBC Persistence-Store is about to be introduced we
should also offer this capability on WildFly / EAP.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)