[rabbitmq-discuss] Network partitioning

[Matthias Radestock]

On 22/03/14 06:02, Geoffrey Samper wrote:
> After disablingquiescing. The problem did not occure this night. If the
> cluster keep running the next couple of days then i can concluded that
> this was a problem and then I will send you the details about how we
> took backups.
> Is this alright?

Certain aspects of virtualisation, such as snapshotting and, apparently, 
quiescing, can cause VMs to effectively stall for a while, which in turn 
would likely be seen as a partitioning event by Erlang. This isn't 
surprising, so I don't think there is much for us to investigate here.


Matthias.