Open Source Database cluster on Windows Server 2003

Question

Hello,

We're looking for an Open Source Database application to run in a cluster environment under Windows Server 2003 across 6-8 cluster nodes. We're primarily looking for High Availability and Load Balancing. Can you provide us with the options available? We've been looking at the SAP MAXDB and Postgres but they're both limited by 1 active instance per cluster which makes the rest of the 7 machines pretty useless! Any suggestions and help would be greatly appreciated.

EDIT: We will be using the database server for an OLTP web application written in PHP. We're looking for high performance with high availability - so we need both replication and automatic failover. I believe we will need a middleware as well to pool the queries since each node should be able to execute write (insert/update) queries as well along with the read queries. So, essentially, I think a setup where the scripts can connect to a middleware and execute the queries - and the middleware can take care of effectively distributing the read/write queries and the replication of database across the nodes -- such a setup would be ideal in my opinion. Is that possible under the Windows operating system with open source technologies?

(We've already tried a number of database servers in the Microsoft Cluster Server environment with a SAN - the only problem being that only one instance at a time is possible)

Thanks.

Answer 1

I don't completely understand your requirements, but...

Slony-I can give you load balancing for read queries.

Pgpool-II and Sequoia should give you statement based middleware (write to all, read from any).

bucardo can do full asynchronous multi-master replication (though it is still growing up -> lot of active development on a main branch).

Have only used bucardo of the above.