You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@activemq.apache.org by Tim Bain <tb...@alumni.duke.edu> on 2016/03/02 18:04:20 UTC

Re: Configuration Questions

1.  The more typical client-server architecture is to have clients connect
directly to the hub broker, but there are valid reasons for using a
client-side broker in a hub-and-spoke topology.  Slow or unstable network
connections are the typical reasons I'm aware of; is that why you've chosen
to use a hub-and-spoke?

2.  This is expected if you're using the default networkTTL value of 1.
Your consumer subscribes to destination D1 on broker0, causing broker0 to
create D1 locally.  Broker0 then subscribes to D1 on every broker it is
directly connected to (networkTTL=1), which includes broker1 but also
broker2-17.  This means that if a producer on any broker publishes to D1,
the message will be forwarded to broker0 and then dispatched to the
consumer.

3.  Sorry, I can't speak to this question, but maybe someone else can.

Tim
On Feb 29, 2016 2:56 PM, "yang.yang.zz" <ya...@outlook.com> wrote:

> Hi:
>
> I have several configuration questions based on the following topology
>
> Here's our *topology*:
>
>                   Consumer0
>                         |
>                     broker0
>
>           /            |              \
>          /      (network)          \
>         /              |                 \
>   broker1        broker2      ...   broker17
>        |               |                      |
>   producer1     producer2   ...  producer17
>
>
> And configuration as below:
>
> *broker0:*
>
>         <transportConnectors>
>             <transportConnector name="openwire" uri="tcp://0.0.0.0:61616
> "/>
>         </transportConnectors>
>
>
> *broker1-17:*
>
>        <networkConnectors>
>                   <networkConnector name="broker0"
> uri="static:(tcp://broker0:61616)" duplex="true" />
>         </networkConnectors>
>       <transportConnectors>
>             <transportConnector name="openwire"
> uri="tcp://
> 0.0.0.0:61616?maximumConnections=100&amp;wireFormat.maxFrameSize=104857600
> "/>
>         </transportConnectors>
>
>
> Questions:
>
> 1)  This is a server-clients topology, is this the right way to configure
> the network connections?
>
> 2)  For the "Queue" menu under broker0's admin web console, I can see the
> Queues from all the client side brokers (broker1-broker17). I makes sense
> as
> these Queues as the communication channel between server and each clients.
> However, at each client side admin web console, I can see the sibling's
> Queues, say at broker1 I can see Queues from broker2, broker3 ... which
> doesn't seem to make sense to me. As broker1 will never communicate with
> its
> siblings. Is this something expected? or just a configuration mistake?
>
> 3) For the "Network" menu under broker0's admin web console, I cannot see
> any "Network Bridges". And from the each client side (broker1-17), I can
> see
> there's a "Network Bridges" connect to broker0. This is different from
> ActiveMQ 5.5.1. In 5.5.1, for the same topology and configuration, from
> broker0, I can see a list of network bridges from all the clients, which is
> very convenient because I can see the number of messages from each clients.
> I noticed this behavior is changed in 5.11.1. Can I know why is this
> changed. Or is this just another configuration mistake?
>
> Thanks a lot in advance for any advice here!
>
> Regards,
> -Yang
>
>
>
>
>
> --
> View this message in context:
> http://activemq.2283324.n4.nabble.com/Configuration-Questions-tp4708596.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>