Some very noob questions on running and configuring

[cmose <cj...@gmail.com>]

Ok, so I appreciate the time of anyone who bears with me on this. I've been
reading through as many of the documents as I can find on activemq and I'm
still scratching my head a little bit on a number of items.

I'm looking at activemq as a standalone, "clusterable" jms "server" to
handle messaging between several servers. Sadly I'm having a hard time
piecing together configuration information, specifically what exactly I need
to do to get this up and running and configured. I see that there is an
example activemq configuration file however, it doesn't strike me as being
teribly clear and there doesn't seem to be a reference for the various tags
and elements listed in the configuration file. Could anyone point me to a
steps 1, 2, and 3 kind of example for getting activemq up and running and
configured? (it's very possible that I've missed that in my perusal of the
documents thus far).

Furthermore, I'm still a bit confused on the use of activemq and jndi, the
application I'm working on is a pojo application and we're actually looking
to get rid of our container altogether (weblogic currently) as we were just
using it for jms. I'm not sure I really follow what needs to happen to use
jndi to lookup an activemq connection factory (I saw the section on spring
but that doesn't really clarify it for me).

And lastly (lord I know, its a lot of questions), it's very possible that my
understanding of spring is not what it should be (although we are going to
try using it, just haven't gotten that far yet), but this quote from the
documentation confused me a bit

"We support full message driven pojo support in Spring by using Jencks. This
provides all the power of MDBs - efficient JMS consumption and pooling of
the message listeners - but without requiring a full EJB container."

I'm not really clear on why jencks would be required to support pojo
messaging - again, perhaps I'm misunderstanding activemq here, but I was
under the impression that activemq could run as a standalone jms provider
and the fact that one was using spring or something else would be
irrelevant, can anyone help me further understand this?

Thanks much, and again, I do really appreciate any insight on this, quite
obviously, simple questions. 
-- 
View this message in context: http://www.nabble.com/Some-very-noob-questions-on-running-and-configuring-tf2078564.html#a5725182
Sent from the ActiveMQ - User forum at Nabble.com.

Re: Some very noob questions on running and configuring

[James Strachan <ja...@gmail.com>]

On 8/9/06, cmose <cj...@gmail.com> wrote:
>
> James,
> thanks very much - I appreciate the information. I know that this is more a
> jencks question then a activemq question, but if wire in a
> ConnectionFactory, and classes each create their own connections and
> sessions, how does that differ from using jencks (at least in those terms) -
> would I be correct in assuming that jencks just abstracts that out?

Yes.

e.g. Jencks will create more sessions & POJOs to deal with largef
message volume. You also have fine grained control over concurrency
(e.g. how many concurrent messages do you want to process for each
subscription) and can share thread pools across different
subscriptions.

Plus there's the transaction handling, XA and retry logic.

But if you are happy with a simplistic threading & pooling strategy,
don't need XA and are happy to write your own transaction handling and
retry code, by all means just use pure JMS yourself.
-- 

James
-------
http://radio.weblogs.com/0112098/

Re: Some very noob questions on running and configuring

[cmose <cj...@gmail.com>]

James, 
thanks very much - I appreciate the information. I know that this is more a
jencks question then a activemq question, but if wire in a
ConnectionFactory, and classes each create their own connections and
sessions, how does that differ from using jencks (at least in those terms) -
would I be correct in assuming that jencks just abstracts that out?
Thanks again!
-- 
View this message in context: http://www.nabble.com/Some-very-noob-questions-on-running-and-configuring-tf2078564.html#a5726683
Sent from the ActiveMQ - User forum at Nabble.com.

Re: Some very noob questions on running and configuring

[James Strachan <ja...@gmail.com>]

On 8/9/06, cmose <cj...@gmail.com> wrote:
>
> Ok, so I appreciate the time of anyone who bears with me on this. I've been
> reading through as many of the documents as I can find on activemq and I'm
> still scratching my head a little bit on a number of items.
>
> I'm looking at activemq as a standalone, "clusterable" jms "server" to
> handle messaging between several servers. Sadly I'm having a hard time
> piecing together configuration information, specifically what exactly I need
> to do to get this up and running and configured. I see that there is an
> example activemq configuration file however, it doesn't strike me as being
> teribly clear and there doesn't seem to be a reference for the various tags
> and elements listed in the configuration file.

See...
http://incubator.apache.org/activemq/xml-reference.html

> Could anyone point me to a
> steps 1, 2, and 3 kind of example for getting activemq up and running and
> configured? (it's very possible that I've missed that in my perusal of the
> documents thus far).

Try

http://incubator.apache.org/activemq/run-broker.html


> Furthermore, I'm still a bit confused on the use of activemq and jndi, the
> application I'm working on is a pojo application and we're actually looking
> to get rid of our container altogether (weblogic currently) as we were just
> using it for jms. I'm not sure I really follow what needs to happen to use
> jndi to lookup an activemq connection factory (I saw the section on spring
> but that doesn't really clarify it for me).

If you are using Spring then there's really no need to use JNDI at all
- just wire stuff together with Spring.


> And lastly (lord I know, its a lot of questions), it's very possible that my
> understanding of spring is not what it should be (although we are going to
> try using it, just haven't gotten that far yet), but this quote from the
> documentation confused me a bit
>
> "We support full message driven pojo support in Spring by using Jencks. This
> provides all the power of MDBs - efficient JMS consumption and pooling of
> the message listeners - but without requiring a full EJB container."
>
> I'm not really clear on why jencks would be required to support pojo
> messaging - again, perhaps I'm misunderstanding activemq here, but I was
> under the impression that activemq could run as a standalone jms provider
> and the fact that one was using spring or something else would be
> irrelevant, can anyone help me further understand this?

The thing jencks gives you is the ability to pool POJOs, connections
and sessions to get efficient JMS consumption together wtih managing
thread pools and dealing with transactions & XA.

If you are happy with a single thread processing the inbound messages,
just create a connection, session and MessageListener and you are
totally fine without using Jencks.

-- 

James
-------
http://radio.weblogs.com/0112098/