Mailing-List: contact users-help@activemq.apache.org; run by ezmlm
Precedence: bulk
Reply-To: users@activemq.apache.org
Received-SPF: pass (nike.apache.org: domain of christian.posta@gmail.com
 designates 209.85.217.176 as permitted sender)
MIME-Version: 1.0
In-Reply-To: 
 <CAJ5znhv9tZfi-MCyn=NniUt1vztHEHGXyRPqhVid8qS6ZQJNxA@mail.gmail.com>
References: <1368173981016-4666877.post@n4.nabble.com>
	<CAKS7T5Ov+ubpmyqNqbL3q8KFqZH5nFTuP2EFcONEz+U8gZ3=6g@mail.gmail.com>
	<1368404686003-4666941.post@n4.nabble.com>
	<CAJ5znhv9tZfi-MCyn=NniUt1vztHEHGXyRPqhVid8qS6ZQJNxA@mail.gmail.com>
Date: Mon, 13 May 2013 07:30:43 -0700
Message-ID: 
 <CAKS7T5PHwCjNow8FW1YdL5UOKrPEKfUtFb75is-Mq3xcsJG03A@mail.gmail.com>
Subject: Re: difficult and unusual configuration..
From: Christian Posta <christian.posta@gmail.com>
To: "users@activemq.apache.org" <users@activemq.apache.org>
Content-Type: multipart/alternative; boundary=089e0160abe0939cb704dc9a5c46

--089e0160abe0939cb704dc9a5c46
Content-Type: text/plain; charset=ISO-8859-1

Paul, that is correct.


On Sun, May 12, 2013 at 7:01 PM, Paul Gale <paul.n.gale@gmail.com> wrote:

> >I saw this document from here
> http://activemq.apache.org/jdbc-master-slave.html and i think in here they
> allow brokers to have more than 2.
> >Did i misunderstand what you mean?
>
> In master-slave you can have as many brokers as you want, regardless of the
> storage strategy in use. However, there can only be one master at any one
> instant in time. The other brokers will remain slaves. As slaves they have
> their transport connectors turned so as not to accept any client
> connections. They are also constantly trying to acquire the shared lock
> that the master hold. If the master dies one of the slaves will be
> successful in its next attempt to acquire the lock as the master will not
> have been able to renew its lease on the lock. Once acquired the slave
> holding the lock will automatically promote itself to become the new master
> at which point it will turn on its transport connectors and start accepting
> client connections.
>
> One requirement for all this to work is that all clients must be using the
> failover protocol to connect to a broker. In turn all clients using the
> failover protocol must know the addresses of all the brokers as
> load-balancing is done client-side; an unfortunate consequence of using the
> failover protocol. I suppose you could mitigate this by putting all the
> brokers behind a load-balancer and have all clients use a single VIP
> although how the load-balancer would know which broker was the new master
> is not clear to me.
>
> At least that's how I've always understood it.
>
> Feel free to correct any errors.
>
> Thanks,
> Paul
>
>
> On Sun, May 12, 2013 at 8:24 PM, verystrongjoe <verystrongjoe@gmail.com
> >wrote:
>
> > thanks for reply.
> > so you mean the master/slave have to configured with only 2 brokers?
> >
> > I saw this document from activemq.org ( here
> > <http://activemq.apache.org/jdbc-master-slave.html>  )
> > and i think in here they allow brokers to have more than 2.
> > did i misunderstand what you mean?
> >
> > and how about using vm connector and putting the jdbc persistent adaptor
> to
> > each broker?
> >
> >
> >
> > --
> > View this message in context:
> >
> http://activemq.2283324.n4.nabble.com/difficult-and-unusual-configuration-tp4666877p4666941.html
> > Sent from the ActiveMQ - User mailing list archive at Nabble.com.
> >
>


-- 
*Christian Posta*
http://www.christianposta.com/blog
twitter: @christianposta

--089e0160abe0939cb704dc9a5c46--