2015-04-28 19:45 GMT+02:00 Erick Erickson <erickerick...@gmail.com>: > I think you're over-thinking the problem though. How often does a > machine fail? If it's more > often than once in an blue moon, you have _other_ problems.
My needs are not only high availability (for which 2 nodes would be enough), but also load balancing. So I wish to have one shard replicated on all three nodes and configure my load balancer (I already have one) to balance requests between three nodes. It can recognise when the service goes down and in this case it removes the failed node from the load balancing pool.