Christoph Gröver Wrote:
-------------------------------------------------------
>
> This sounds to me as if it is done by MySQL
> automatically ?
> So, if you have 5 times more Storage Nodes than
> replicas
> (replica means SQL Node, right ?) each node has
> only to
> store 1/5th of the data ?
Replicas has nothing to do with the number of SQL nodes connected to the cluster, you can have as many as you requre. A replica, either 1 2 or 3 is the number of full copies of the database stored within your Data (NDBD) nodes. Recommended config is 2, so that you can handle node failures etc, and 3 is available but extreme in most cases.
Based on that, if you have replicas configured as 2. and have 16 datanodes, you will have 2 copies of the database, 1 copy stored accross 8 of the datanodes, and the second copy stored on the other 8. MySQL handles this process automatically.
This page in the manual makes soem of the terms clear and has a diagram on what I just discussed.
http://dev.mysql.com/doc/refman/5.1/en/mysql-cluster-nodes-groups.html