[rabbitmq-discuss] Recover cluster from crash

[Marek Stój]

Turns out the problem is that the environment variable RABBITMQ_NODE_ONLY 
is not taken into account on Windows version of RabbitMQ server (sic!). One 
can workaround this by manually removing ''-s rabbit boot ^" line from 
rabbitmq-server.bat.

On Monday, April 14, 2014 3:04:30 PM UTC+2, js.future... at gmail.com wrote:
>
> Hello,
>
> we are testing RabbitMQ for future replace MSMQ. Someday ago we had a 
> crash with one node. Cluster is build with 3 nodes HA on Windows machines - 
> RabbitMQ instances are services. This node has been reinstalled. After this 
> cluster failed to start. Now on other two machines we get information that 
> can not connect to other nodes. We tried set enviroment variable RABBITMQ_NODE_ONLY, 
> run service and executed command to remove node "rabbitmqctl 
> forget_cluster_node --offline xxx at nodeName". On running this command we get 
> information that node is not found "Error: {not_a_cluster_node,"The node 
> selected is not in the cluster."}" what to do with this? We don't want to 
> reinstall all. If we start it normally we get exception in log that:
>
> *"*
>
> *Error description:   {error,{inconsistent_cluster,"Node rabbit at X thinks 
> it's clustered with node rabbit at Y, but rabbit at Y..."}}*
> *"*
>
> Please help us.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20140425/b195b673/attachment.html>