[rabbitmq-discuss] Cluster reverted to old config!

Simon MacMullen simon at rabbitmq.com
Fri Apr 5 12:11:36 BST 2013

On 04/04/13 13:06, Pete Kelly wrote:
> Any advice or thoughts would be appreciated, if only to help prevent
> this from happening in the future!

To add to what Emile already said: RabbitMQ prior to 2.8.0 did not force 
an fsync on declare / delete operations for durable objects, so it was 
possible for such operations to get (effectively) rolled back after a 
crash. I don't know how long you can manage to go without an fsync 
though - I would expect you'd struggle to get to an hour and it sounds 
like your definitions were much older than that.

However, an upgrade is, as usual, recommended :-)

Cheers, Simon

Simon MacMullen
RabbitMQ, VMware

More information about the rabbitmq-discuss mailing list