You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@activemq.apache.org by dhurlburtusa <da...@realgo.com> on 2006/06/03 00:33:36 UTC

Pop Quiz. Who can answer these questions correctly and throughly?

I can't find any useful documentation. All I find is explanations about a
bunch of features ActiveMQ has but nothing says how to use it. The current
documentation is like a dictionary. It explains/defines all the words but it
does not tell you how to put them together to compose an essay. I don't mean
to knock the documentation but it really needs some work. I would be happy
to contribute but I can't if I can't understand how to use ActiveMQ.

Some things I have questions about:

What is a transport connector? How is it used? How do I configure one?

What is a network connector? How is it used? How do I configure one?

What is the difference between a transport connector and a network
connector?

Where can I find a full DTD to use in my XML configuration file? All I find
is a partial schema or DTD. Neither have definitions to the elements defined
by Spring's configuration.

What happens to a slave broker after it becomes a master and the original
master is restarted? There is no documentation about what happens during
recovery. When the original master restarts, does the current master (which
was the previous slave) go back into slave mode? Do all the clients
reconnect to the newly restarted master?

The current documentation says that a master/slave broker configuration is
recommended. Why? What are its advantages and disadvantages?

Can you have a master/slave/slave/... setup? For example: use
(failover://tcp://masterhost:61616,tcp://slave1host:61616,tcp://slave2host:61616tcp://slave3host:61616)?randomize=false

Can you setup a network of master/slave brokers? Would this even be useful?

About the example broker configuration files. Why is that not a <!DOCTYPE or
a xml namespace and schemaLocation? None of my XML editors validate the
files. None of the editors provide auto-completion with out a proper DOCTYPE
or schema reference.     

After looking at the  xsd  I am confused. Many examples I see have
networkConnectors elements and transportConnectors elements. All I find is
networkConnector or transportConnector elements in the XSD. Where is the
documentation explaining the purpose of the attributes for each element?

Help with these questions might allow me to keep my current job.

Thanks

--
View this message in context: http://www.nabble.com/Pop-Quiz.-Who-can-answer-these-questions-correctly-and-throughly--t1725429.html#a4688061
Sent from the ActiveMQ - User forum at Nabble.com.


Re: Pop Quiz. Who can answer these questions correctly and throughly?

Posted by "Christopher G. Stach II" <cg...@ldsys.net>.
James Strachan wrote:
> On 6/2/06, dhurlburtusa <da...@realgo.com> wrote:
>> Can you setup a network of master/slave brokers? Would this even be
>> useful?
> 
> Yes and yes. To get store/foward across master/slave pairs.

I asked in another thread, but I'll ask again.  Can we see an example of
this configuration?

> The generated XSD is not perfect yet but it should be a start if you
> wanna use it. Note that we can't use DOCTYPEs as we are using XML
> namespaces; so its XSDs only I'm afraid.

That's a good thing. :)

-- 
Christopher G. Stach II

Re: Pop Quiz. Who can answer these questions correctly and throughly?

Posted by James Strachan <ja...@gmail.com>.
On 6/2/06, dhurlburtusa <da...@realgo.com> wrote:
>
> I can't find any useful documentation. All I find is explanations about a
> bunch of features ActiveMQ has but nothing says how to use it. The current
> documentation is like a dictionary. It explains/defines all the words but it
> does not tell you how to put them together to compose an essay. I don't mean
> to knock the documentation but it really needs some work. I would be happy
> to contribute but I can't if I can't understand how to use ActiveMQ.

Am looking forward to your contributions :)


> Some things I have questions about:
>
> What is a transport connector? How is it used? How do I configure one?

Its used to connect a broker to clients.


> What is a network connector? How is it used? How do I configure one?

Its used to connect a broker to other brokers.


> What is the difference between a transport connector and a network
> connector?

One talks to clients, one talks to other brokers (in a network or
master/slave etc)


> Where can I find a full DTD to use in my XML configuration file? All I find
> is a partial schema or DTD. Neither have definitions to the elements defined
> by Spring's configuration.

http://activemq.org/site/xml-reference.html

unfortunately the code generation is a bit lacking. The simplest way
to figure stuff out is often to look at the BrokerService which is the
POJO which is reflected to configure a broker.


> What happens to a slave broker after it becomes a master and the original
> master is restarted?

Currently the slave remains a master and its a manual process to bring
the old master back online (by copying the files from the new master
to the old master).


> Do all the clients
> reconnect to the newly restarted master?

Clients only reconnect when the broker they are connected to goes down.


> The current documentation says that a master/slave broker configuration is
> recommended. Why? What are its advantages and disadvantages?

Networks are designed for store-forward topologies really, not for
master/slave. i.e. in a network a message moves from one broker to
another. With master/slave the messages are replicated to 2 brokers.


> Can you have a master/slave/slave/... setup?

Not yet no.


> Can you setup a network of master/slave brokers? Would this even be useful?

Yes and yes. To get store/foward across master/slave pairs.


> About the example broker configuration files. Why is that not a <!DOCTYPE or
> a xml namespace and schemaLocation? None of my XML editors validate the
> files. None of the editors provide auto-completion with out a proper DOCTYPE
> or schema reference.

The generated XSD is not perfect yet but it should be a start if you
wanna use it. Note that we can't use DOCTYPEs as we are using XML
namespaces; so its XSDs only I'm afraid.

We generally don't mandate the XSD in each XML config file but leave
it as an option for the end user to use if they wish (validating XML
with XSDs often causes problems with XML parser classpaths or the
finding of the XSDs - the main value of XSDs seems to be for editor
tooling so use it if you like)

-- 

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

Re: Pop Quiz. Who can answer these questions correctly and throughly?

Posted by "Christopher G. Stach II" <cg...@ldsys.net>.
dhurlburtusa wrote:
> What is a transport connector? How is it used? How do I configure one?

Consider it the broker's inbound service.

http://activemq.org/site/xml-reference.html

> What is a network connector? How is it used? How do I configure one?

Consider it a connection out of the broker to other brokers.

http://activemq.org/site/xml-reference.html

> What is the difference between a transport connector and a network
> connector?

See above.

> Where can I find a full DTD to use in my XML configuration file? All I find
> is a partial schema or DTD. Neither have definitions to the elements defined
> by Spring's configuration.

How about XSD?  See above.

> What happens to a slave broker after it becomes a master and the original
> master is restarted? There is no documentation about what happens during
> recovery. When the original master restarts, does the current master (which
> was the previous slave) go back into slave mode? Do all the clients
> reconnect to the newly restarted master?

Nothing.  You're hosed until you restart both.  There's no recovery
other than manual intervention.  Reconnection is up to the clients.

> The current documentation says that a master/slave broker configuration is
> recommended. Why? What are its advantages and disadvantages?

I don't know.  I think it's useless since it doesn't support recovery.
I'd rather use (and do use) discovery or static network connectors
between three or so machines, use journaled persistence, and XA.  If you
don't want to lose any messages, do it the right way, even if it's
slower than async and not persistent.

> Can you setup a network of master/slave brokers? Would this even be useful?

I don't think so.  I asked and got no reply.  Next peer replication
would be best.

-- 
Christopher G. Stach II