Sorry this is a bit late. I have some comments & questions on the EJB,
JMS, and JCA alignment proposal.
General comments
* For consistency with the platform spec, the lookup strings should be
"java:comp", not "java:/comp".
* JNDI names should be in "java:comp", not "java:comp/env".
java:comp/env is owned by the app. Using env could cause backward
compatibility issues for apps that have declared these names.
* JNDI names should be capitalized for consistency with other builtins:
UserTransaction, TransactionSynchronizationRegistry, ORB, EJBContext,
TimerService, etc.
Section 2.2.2
* uniqueMDBName - What is the scope of the uniqueness? Within the current
module, app, server, cluster, or UUID?
* instanceName - This name is confusing in the context of an EJB. Perhaps
"ServerInstanceName" or "ClusterMemberName"?
Section 2.2.3
* inAppClientContainer - Why was type String chosen instead of Boolean? (I
think this may just be a copy/paste bug)
* inAppClientContainer -Since this property is component specific it seems
to make more sense to bind it into java:comp. (In addition, using shared
namespaces such as java:global or java:app to store this setting will
cause complications for some of the federated deployment modes that
WebSphere supports.)
General questions
* Do "java:comp" entries need to be available outside the call to
activateEndpoint? If yes, what are the semantics for EJB-in-WAR (wherein
all components share java:comp)?
* Further, do the "java:comp" entries need to be available when
activateEndpoint is called outside the scope of an MDB?
-Jeremy
From: Marina Vatkina <marina.vatkina_at_oracle.com>
To: jsr345-experts_at_ejb-spec.java.net,
Date: 07/05/2012 08:11 PM
Subject: [jsr345-experts] Fwd: Proposals for MDB and RA
standardisation (JMS_SPEC-30,55,54,25,73)
Experts,
This is the email and the revised proposal on alignment of EJB, JMS, and
JCA specs in the MDB configuration areas.
Please read carefully the email and the proposed changes and let me and
Nigel (who is on the users alias) know if you agree with the proposal or
not (and if not, the why).
thanks,
-marina
-------- Original Message --------
Subject: [jms-spec users] [jsr343-experts] Proposals for
MDB and RA
standardisation (JMS_SPEC-30,55,54,25,73)
Date: Fri, 29 Jun 2012 18:59:23 +0100
From: Nigel Deakin <nigel.deakin_at_oracle.com>
Reply-To: jsr343-experts_at_jms-spec.java.net
To: jsr343-experts_at_jms-spec.java.net
This email covers a number of related issues, all covered in the attached
document.
Some time ago we discussed and agreed on a number of changes that we would
like to see made to the EJB spec to
standardise the way in which JMS MDBs were configured. Here are the JIRA
issues:
http://java.net/jira/browse/JMS_SPEC-30
Define mandatory activation config properties for JMS MDBs
http://java.net/jira/browse/JMS_SPEC-55
Define a standard way to configure the connection factory used by a JMS
MDB to consume messages
http://java.net/jira/browse/JMS_SPEC-54
Define a standard way to configure the destination on which a JMS MDB
consumes messages
I have spent quite a lot of time negotiating with the EJB spec lead about
the details of these changes and have now come
to provisional agreement, though these change have yet to be confirmed by
the EJB EG. The attached document contains a
summary of what we agreed. I would now like to bring these proposals back
to the JMS EG for comments.
Separately, also agreed in principle even longer ago that we should
standardise the interface between a JMS provider and
a Java EE application server by making it mandatory for a JMS vendor to
provide a JCA resource adapter. Here is this
JIRA issue:
http://java.net/jira/browse/JMS_SPEC-25
Standardise the interface between a JMS provider and a Java EE application
server
However if we're going to standardise on MDB configuration we also need to
standardise on ActivationSpec properties. The
attached document also contains some proposals to define a JMS resource
adapter as part of the JMS specification. This
essentially consists of moving the section on JMS resource adapter out of
the JCA 1.6 spec, adding it to the JMS spec,
and adding some additional properties.
Finally, on 15th June I raised this new issue:
http://java.net/jira/browse/JMS_SPEC-73
Define how messages from a topic are delivered to clustered application
server instances
I made some provisional proposals to address this here
http://java.net/projects/jms-spec/lists/jsr343-experts/archive/2012-06/message/7
I haven't received any comments on that email. However since these
proposals involve the standardisation of an
additional ActivationSpec property, and the container making varius pieces
of information available to the resource
adapter, I thought it helpful to combine these proposals within this
document as well.
So please have a look at the attached document and make any comments. I'm
not setting a deadline for comments but it
would be really helpful if you could have a look within the next week or
so. If it's not obvious why I've made a
particular proposal please do ask and I'll try to explain.
I know we're approaching the vacation season, and it's the 4th July next
Wednesday as well.
If this EG is generally happy with this document I plan to turn this
document (except for the parts that go into the EJB
spec) into a new chapter for the JMS spec.
Thanks,
Nigel
[attachment "MDBAndRAConfiguration.pdf" deleted by Jeremy
Bauer/Rochester/IBM]