[rabbitmq-discuss] Repairing a crashed cluster - Obtaining persisted msgs from master

Simon MacMullen simon at rabbitmq.com
Wed May 1 10:58:14 BST 2013

On 30/04/13 18:10, ranjitiyer wrote:
> Hello,
> Part of my question was answered here
> http://rabbitmq.1065348.n5.nabble.com/Repairing-a-a-crashed-cluster-td22466.html.
> I have a follow up question for the following scenario.
> Say, I have a master/slave cluster with mirrored persisted queues. Slave
> dies, messages m1, m2 are published to the master. Master dies. Slave comes
> up - it can't because master is dead. I realized am going to have to follow
> the instructions on the previous post to get erl.exe and eventually the
> rabbit running on the slave machine.

Note that since that message, RabbitMQ 3.0 was released, with 
"rabbitmqctl forget_cluster_node" - which might help you get the slave 
started if the master is gone.

> Good so far. I now also want the messages of the master to be brought over
> to the slave because I don't wanna lose m1 and m2 sitting on the master
> machine. I know that queue messages are persisted in the mnesia database
> located under c:\user\appdata\...\db\. Is there a way to export data from
> this db and import it into the db on the slave machine?

So the Mnesia database itself can't be moved - but the subdirectories 
"queues" "msg_store_transient" and "msg_store_persistent" probably can. 
You could try that - I would take backups of everything first though.

But what are you trying to do? If the master is still available and more 
up to date than the slave then why are you removing it?

Cheers, Simon

Simon MacMullen
RabbitMQ, VMware

More information about the rabbitmq-discuss mailing list