Distribution of Replicas among nodes

Luke Bakken lbakken at basho.com
Thu Feb 11 15:20:32 EST 2016

Hi Gerhard,

I believe setting up your cluster that way gives the greatest
probability that your data will exist on every node.

Someone with more knowledge of the hashing method could chime in about


On Tue, Feb 9, 2016 at 12:17 AM, Pauls, Gerhard <Gerhard.pauls at atos.net> wrote:
> Hi all,
> i want to setup a Riak DB such that:
> all nodes have one replica of the data, so that any single node can work on
> it’s own in case of network failures.
> I’m intending to reach that goal by:
> Setting n_val equal number of nodes, e.g. 4
> Making sure that the Ring Size (e.g. 64) can be divided evenly by the number
> of nodes
> Is it correct that in this case all the data will always be distributed
> evenly on all the nodes, that means each node will have a copy of all the
> data?
> Kind regards,
> Gerhard

More information about the riak-users mailing list