[rabbitmq-discuss] Mnesia corrupting after node joining cluster

[David Brown]

Matthias,

thanks so much for your info on this.  I assumed from the start that I 
didn't need to cluster, but I never ran across any documentation that 
referred to using a single broker, the '...distributed.html' was misleading 
and the 'RabbitMQ In Action' book never pointed this out.

I'm developing using the .Net client in C#, this is my method to connect to 
rabbit:


   public bool connectToRabbitMQ()

    {

        bool r = true;

        try

        {

            var connectionFactory = new ConnectionFactory();

            connectionFactory.HostName = mServer;

            mConnection = connectionFactory.CreateConnection();

            mModel = mConnection.CreateModel();

            mModel.BasicQos(0, 1, false);

            mExchanges = new List<string>();

        }

        catch (BrokerUnreachableException e)

        {

            r = false;

        }



        return r;

    }



So all I need to do is have the variable mServer contain the name of the 
Host that is running the RabbitMQ broker, right?



Thanks again



David







----- Original Message ----- 
From: "Matthias Radestock" <matthias at rabbitmq.com>
To: "David Brown" <dbrown at prmllc.com>; "Discussions about RabbitMQ" 
<rabbitmq-discuss at lists.rabbitmq.com>
Sent: Wednesday, August 01, 2012 2:23 PM
Subject: Re: [rabbitmq-discuss] Mnesia corrupting after node joining cluster


> David,
>
> On 01/08/12 19:15, David Brown wrote:
>> To implement distributed messaging, my understanding is I've got three
>> options, cluster, federation, shovel.
>
> I guess the title of http://www.rabbitmq.com/distributed.html might have 
> given you the impression that to get a message flowing from an app on one 
> machine to an app on another machine you'd need brokers on both machines 
> and connect them with clustering, federation, or shovel.
>
> That is not the case.
>
> AMQP is inherently distributed - it is a network protocol, so it's quite 
> typical to have a single broker to which apps connect from wherever they 
> are running. Apps and brokers do *not* have to reside on the same machine 
> (and, typically don't).
>
>
> Regards,
>
> Matthias.
>
>