[rabbitmq-discuss] How to create a rabbitmq cluster using existing application without losing existing data?

[Joshua Lim]

> Date: Thu, 12 Dec 2013 16:40:46 +0000> From: simon at rabbitmq.com
> To: joshua__lim at hotmail.com
> CC: rabbitmq-discuss at lists.rabbitmq.com
> Subject: Re: [rabbitmq-discuss] How to create a rabbitmq cluster using existing application without losing existing data?
> 
> On 12/12/2013 16:34, Joshua Lim wrote:
> > How about the cluster_nodes configuration field on existing machine?  I 
> > left it as {cluster_nodes, {['rabbit at existing-machine', 
> > 'rabbit at new-machine'], disc}}
> 
> That field will only even be read when the node in question comes up
> with a completely blank database (i.e. first start, or after
> 'rabbitmqctl reset'). That presumably never applies to the existing
> machine, so it doesn't matter what you have there.
> 
> Cheers, Simon


Thanks.  I tried redoing the whole test from scratch - resetting everything new-machine.
When I tried to cluster rabbit at new-machine with rabbit at existing-machine by adding the cluster auto config to new-machine:
       	         {cluster_nodes, {['rabbit at existing-machine'], disc}}
I get the following error:
Error description:   {error,{inconsistent_cluster,"Node rabbit at new-machine thinks it's clustered with node rabbit at existing-machine, but rabbit at existing-machine disagrees"}}
I managed to solve this by adding the same auto config to rabbit at existing-machine - this must be done before rabbitmq is started for the first time - which means I need to reset my existing machine.  Is this correct, or am I missing something?
       	         {cluster_nodes, {['rabbit at existing-machine'], disc}}
Rgds,Joshua





 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20131213/17112b77/attachment.html>