[rabbitmq-discuss] How do i stop the erlang node but keep the rabbitmq server running.

[Prashanth M]

Hi Tim,

Yes, I am trying to see what happens if you've got two nodes and one goes
down. Clustering doc describes about the cluster setup, it does not tell
how to test the cluster
I have two nodes, when I bring down a node with rabbitmqctl stop_app, my
web app stops working. I am not sure what I am doing wrong. But I can see
that the queues are duplicated between two nodes. my web app pointing to
the stomp on the node that is brought down.

Thanks


On Fri, Jul 6, 2012 at 2:06 PM, Tim Watson <tim at rabbitmq.com> wrote:

> Hi Prashanth,
>
> Please keep replies on list, as I mentioned earlier (just before
> forgetting to do some myself!). :)
>
> On 06/07/2012 18:38, Prashanth M wrote:
>
>> correct me if i am wrong. We proved the reliability of rabbitmq, we are
>> planning to prove how clustering works out of box.
>>
>>
> Ok that's fine, but *what* about clustering do you need to prove for your
> architecture? Are you wanting to know what happens if you've got two nodes
> and one goes down? The idea is that all the relevant parts that should be
> present on all nodes, should survive a crash. As the list entry I referred
> you to in my previous mail states, you've got to be careful when using ram
> nodes, as you *can* get yourself into a situation where data loss is
> *possible* under those circumstances.
>
> Please read the clustering guide in detail and then feel free to come back
> ask any questions. If you want to stop rabbit on a node, without stopping
> the node itself, you'd use `rabbitmqctl stop_app` or `rabbitmqctl stop` to
> stop the rabbit application *and* the node.
>
> Tim
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20120709/65122191/attachment.htm>