Active\Active Shared Database.

Thanks shea. The shared disk will not be accessed by more than one node. Each node will have its own shared bay and database. But the data will be federated\partitioned across the databases on both instances horizontally. Using DP Views the application ac
cess the data required. This is my thought and not sure whether practically possible or not. This way we can use the resources of both the nodes (scale out). When a failover happens, a single node will share the load which is obvious in case of Oracle's R
eal Application Cluster also.
Marx.
You can create 2 SQL Instances. Each partitioned table will reside in a database in each SQL Instance. The distributed view will read/write data to the correct SQL Instance. If SQLInstance1 is on Node1 and SQLInstance2 is on Node2 you will achieve a de
gree of load balancing via the distributed views. However, if Node1 fails SQLInstance1 will failover to Node2 and you will now have degraded performance. If you wish to maintain this high availability AND high performance scenario then you need 2 x 2-no
de clusters. Then if any node fails in either cluster you will maintain the same level of performance even though you now have a reduced level of redundancy. Of course this solution costs a whole lot more...
When ever you design a clustered solution you need to ask your clients if they are prepared to take a reduced level of service (in this case performance) when you loose some of your redundancy (due to a hardware or software failure).
|||That will work. Managing the DPV is challenging to say the least, but it is
possible.
Mike
Principal Mentor
Solid Quality Learning
"More than just Training"
SQL Server MVP
http://www.solidqualitylearning.com
http://www.mssqlserver.com