From: Craig L Russell
Date: November 28 2012 7:20pm
Subject: Re: Cluster J scalability
List-Archive: http://lists.mysql.com/cluster/8452
Message-Id: <08FD56B1-783F-4612-B817-9D6FB17DBA00@oracle.com>
MIME-Version: 1.0 (Apple Message framework v936)
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit

Hi Graeme,

The connect string does point to the management node, which is only  
used to establish the connection. Thereafter, management node doesn't  
do anything active with regard to sessions; the client makes a direct  
tcp connection to each data node in the cluster. All the data traffic  
is between client and data nodes.

Clusterj supports session pooling by changing the connection properties.

Take a look at clusterj Constants and in particular the property  
PROPERTY_CONNECTION_POOL_SIZE.

You can find code examples in clusterj-test ConnectionPoolTest.java.

Regards,

Craig

On Nov 28, 2012, at 11:13 AM, Graeme Wallace wrote:

> When we're using JDBC to connect to MySQL cluster - we have to have
> multiple query nodes to spread the load.
>
> However, ClusterJ is configured by setting up a connect string.  
> Currently I
> have this pointing to my management node. How scaleable does this  
> end up
> being ? ie How much resource is the management node having to  
> provide in
> order for a query/insert/update etc ? Do i need to think about having
> multiple SessionFactory's in my code to connect to different nodes to
> spread the load ?
>
>
> -- 
> Graeme Wallace
> CTO
> FareCompare.com
> O: 972 588 1414
> M: 214 681 9018

Craig L Russell
Architect, Oracle
http://db.apache.org/jdo
408 276-5638 mailto:Craig.Russell@stripped
P.S. A good JDO? O, Gasp!