The first option, the coordinator node takes care of sending the work to the other nodes. It will return to you when the write has been acknowledged by the number of nodes you specify in the consistency level. 


On 07 Jul, 2010,at 06:06 PM, ChingShen <> wrote:

Thanks aaron morton,

   I have an another question about replication. if I run a write operation on A node and how does the replication work?  A node dispatches three request to B, C and D nodes respectively to do write operation? or A node dispatches a request to B node, then B node forwards it to C node, finally, C node forwards it to D node?



On Tue, Jul 6, 2010 at 8:04 PM, aaron morton <> wrote:
Which ever node the client connects to is the coordinator node. It will take care of sending the messages to the nodes who will do the actual work.

If you have multiple nodes try a DNS round robin to distribute client connections around.

On 6 Jul 2010, at 22:10, ChingShen wrote:

> Hi all,
>   I'm newbie in Cassandra, I have a question about which node is coordinator node in my cluster.
> I have A, B, C, D, E, F and G nodes. if I run a write operation on "A" node, and key range between A and B, so the A node is responsible to write the key to B, C and D nodes(RF=3) ? and does it mean A node is coordinator node?
> Thanks.
> Shen