Now, I am using 1 replication with local gateway but I am suffering from
writing performance gradation while shard relocation and server restart.
I can mitigate this problem if I increase the size of the cluster and
prevent the cluster interruption as possible but I want to prepare the
worst case.
So, I want to try no replication with Amazon EBS volume. If you have tried
this option, I will appreciate so much if you share your experience.
Thank you
Best, Jae
--