incubator-cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ran Tavory <ran...@gmail.com>
Subject Re: Client connection and data distribution across nodes
Date Thu, 17 Jun 2010 06:07:03 GMT
On Thu, Jun 17, 2010 at 8:52 AM, Mubarak Seyed <seyed@apple.com> wrote:

> Hi All,
>
> Regarding client thrift connection, i have 4 nodes which formed a ring, but
> client only knows the IP address of an one node (and thrift RPC port
> number),
> how does client can connect to any one other node without getting ring
> information? Can we keep the load balancer and bind all the four nodes or
> client needs to know the IP address of all the 4 nodes?
>
> If you use java there are higher level libraries that manage ring
information for you, so they may help. If not, I guess you'll need to call
the describe_ring thrift api.


> Regarding storage management, for instance, if we want to store 100k
> records, but each 25k records on each node, something like
>
> node 1 -> 25K
> node 2 -> 25K
> node 3 -> 25K
> node 4 -> 25K
>
> Can we accomplish using OrderPreservingPartitioner (OPP)? How does
> replication happen between nodes if we keep only 25k records in one node?
>
> Can someone please let me know. Thanks in advance.
>
> Thanks,
> Mubarak
>
>
>
>
>
>

Mime
View raw message