?? 2010-06-01???? 15:00 +0800??huajun qi??????
> If you have a multinode cluster, which node you should connect to
> fetch data?
any one.
>
> Is there a master node in a cluster which accepts data request and
> dispatch it? Or every node in the cluster is completely same?
no master. all the same.
>
> If all nodes are same in a cluster, should client connect to random
> node to reduce cassandra's load?
I think so. But I guess if you're sure where the data is, you can
connect the target machine directly.
> --
> Location:
>
Kevin Yuan
